An Introduction to Indonesian Linguistics/Essay 2

by Renward Brandstetter

ESSAY II

COMMON INDONESIAN AND

ORIGINAL INDONESIAN

(The original was issued as an Appendix to the Annual Report of the Cantonal School, Lucerne, 1911.)

SUMMARY

1-16. Introduction.

17-177. Part I : Common Indonesian.

17-49. Section I : Phonetic System. 17-27. Vowels. 2s 32. Diphthongs. 33-48. Consonants. 49. Table of Sounds.

50-74. Section II : Synthesis of Sounds into Words. 50. Disyllables. 51-2. Monosyllables. 53. Trisyllables. 54. Initial Sounds. 55-60. Medial Sounds. 61-73. Final Sounds. 74. General Type of the Word.

75-80. Section III : Accent.

81-142. Section IV: Formal Analysis of Words. 81. Preliminary Observations. 82-3. Interjections. 84-114. Words of Form. 84. General Characteristics. 85-95. Articles. 96-106. Prepositions. 107-9. The Particle a. 110-3. Negatives. 114. Conjunctions. 115-20. Words of Substance. 115. General Characteristics. 116-7. Roots. 118-20. Words. 121-35. Pronouns. 121-3. General Characteristics. 124-31. Personal Pronouns. 132. Demonstrative Pronouns. 133-4. Interrogative Pronouns. 135. The Indefinite Pronoun. 136-42. Numerals.

143-73. Section V: Extension of the Word-base. 143-6. Preliminary Observations. 147-61. The Verb. 147. Verbal Formatives. 148-51. Active Formatives 152-4. Passive Formatives. 155. Their force. 156. The Transitive Formative -i. 157. The Causativ Formative pa-. 158. Common Indonesian Verbs with the same Formative. 159-60. The Imperative without a Formative. 161. Formatives for Tenses. 62-7. The Substantive. 162. Substantives without Formatives. 163-7. Substantival Formatives. 168-70. The Adjective. 168-9. The Adjectival Formative ma-. 170. The comparative Formative. 171-2. The Adverb. 171. Adverbs without Special Formatives. 172. The Adverbial Formative ka- . 173. The Numeral

174-6. Section VI : Reduplication of the Word-base. 177. Synthesis of Words into Sentences. 178-91. Part II : Original Indonesian.

INTRODUCTION.

1. In the Tagalog language of the Philippines the word for " sky " is laṅit, and it has also the same form in the Tontemboan of Celebes, the Dayak of Borneo, the Javanese of Java, the Gayo of Sumatra, the Malay of the Malay Peninsula, the language of the Mentaway Islands (which lie to the south-west of Sumatra), and in many other Indonesian languages besides these. It is true that in the language of the Batan Islands, northward of the Philippines, we find gañit, in the Bimanese of Sumbawa, an island lying towards New Guinea, lañi, and in the Hova of Madagascar lanitra; but it can be proved by means of strict phonetic laws that these three tongues, in an earlier stage of their development, also used the form laṅit. — We have, therefore, in many IN ( = Indonesian) languages one and the same expression for " sky ", viz. laṅit.

Note. —The accentuation of IN words, including therefore laṅit, is dealt with in § 75.

2. Such IN linguistic material as recurs in many different languages either unchanged or modified only in accordance with strict phonetic laws, we style Common Indonesian. We say, therefore, that there is a Common IN name for the sky, viz. laṅit.

3. The wider the distribution of an IN linguistic phenomenon, the more positively shall we be entitled to pronounce it to be Common IN. Our right to do so will be particularly strong when the phenomenon manifests itself at the most different points of the IN linguistic area: i.e., to put the matter concretely, in the seven great insular regions and the three border districts.

Note I. — The seven great insular regions are: the Philippines, Celebes, Borneo, Java with Madura and Bah, Sumatra, the Malay Peninsula with the adjacent islands, and lastly Madagascar. — The three border districts are : in the North, the Batan Islands and Formosa; in the East, the islands from Lombok towards New Guinea, of whose languages the best known to us are Bimanese, Kamberese, Sawunese, Rottinese, Tettum, and Masaretese; in the South -West, the row of islands behind Sumatra: Simalur, Nias and Mentaway. — Another border district, viz. Halmahera * and the adjacent islands, cannot be included in our survey because there is a doubt whether it really belongs to the IN linguistic area: see Van Hinloopen Labberton, " Handboek van Insulinde ", p. 88.

Note II. — It is of course not to be expected, and indeed it will seldom happen, that we shall be able to demonstrate the existence of one and the same linguistic phenomenon in all these ten areas of distribution; the Malay Peninsula area in particular, with its impoverished vernacular, will often fail us. Our normal standard will be nine, eight, or it may be only seven areas: if the number is less than that, it will -be only with diffidence, if at all, that we shall pronounce a linguistic phenomenon to be Common IN.

4. When we confront together languages thus diversely situated in geographical position, we are at the same time comparing languages that are related in the most various degrees of relationship. It will happen, for instance, that we shall compare the Malay of the Peninsula, the Minangkabau, of Sumatra, the Javanese, and the Masaretese of the eastern ' border. Now the Minangkabau is closely allied to the Malay, ' the Javanese is more distantly connected, the Masaretese still more distantly; and the same sort of thing will occur in all our comparisons. Now if we find one and the same linguistic phenomenon occurring in several forms of speech which are related to one another in the most various degrees, that is to say, even in such as in other respects are most distant relatives, then we can with perfect confidence regard such a phenomenon as being Common IN.

* [This applies in particular to the northern part of Halmahera: see Huetuig's article in Bijdr. 1907-8, pp. 370 seqqq]

5. The word for "cloud " in the Pampanga of the Philippines is biga, in the Bugis of Celebes ĕllun, in the Malay of the Peninsula awan : in short, one might almost say that each IN language has its own special word for “cloud”. We therefore say: there is no Common IN expression for “cloud” .

6. Now the Common IN linguistic phenomena form the subject of the first of the two prmcipal -parts into which our monograph is divided.

7. Let us now take stock of the chief principles of method which must serve as our lodestars in this first part. Herein we must first of all realize that our delineation of the Common IN element must have two facets, a positive and a negative side. If we establish the fact that there are Common IN expressions for “sky” , “to weave”, and “ten”, we must also at once add that for the concepts “cloud”, “to spin”, and “eleven” there are no Common IN designations. If we gave the reader only the positive results, our sketch would be one-sided, partial, and too favourable.

8. Further, we must build up our demonstrations entirely upon the basis of phonetic laws. That is really self-evident, bat it must nevertheless be particularly insisted on here, because this first part of the monograph has to yield us a thoroughly sound foundation for the second part, which is of a more hypothetical nature, and therefore less certain.

We must not therefore content ourselves with maintaining, for example, that Bimanese laṅi and Hova lanitra are derived from an original form laṅit : we must formulate the laws in accordance with which laṅi and lanitra have come into being. These laws are as follows: “In Bimanese and Sawunese all Common IN final consonants disappear”, therefore laṅi < laṅit.<r1> — “Every Common IN ṅ becomes n in Hova, except before a velar” " ; “every Common IN final t appears in Hova as -tra” ; therefore lanitra < laṅit. — As to Batanese gañit, see § 9.

9. Besides having recourse to phonetic laws we shall find references to parallel cases of great service. In Mentaway laṅit no longer means “sky”, as it does in Common IN, but “the red tint of dawn and sunset”. This transition in meaning would hardly, I imagine, disconcert us: instead of the sky we have a phenomenon in the sky. But our confidence will be even greater when we observe that a parallel case occurs in a dialect of Formosa, where araṅit < laṅit means “cloud” : here, too. instead of the sky we have a phenomenon in the sky.

It is, however, not only in our researches into the varying significations of words, that this method of reference to parallel cases will assist us: we shall also occasionally apply it instead of the appeal to phonetic laws. As stated in § 1, Common IN laṅit, “sky”, appears in Batanese under the form gaṅit. Now the hitherto published Batanese material includes barely a hundred words, and among these there are only three cases in which ñ represents Common IN ñ. Three cases are, however, too few to enable a phonetic law to be formulated with safety. Here, therefore, we take refuge in a parallel and say : In Batanese gaṅit < laṅit, ñ represents Common IN ñ, as in aṅin < Common IN aṅin, “wind”.—Further, among those hundred words, the number of cases in which Batanese g represents Common IN l is somewhat larger, there are ten safe cases ; but even that number seems to me too small to enable a phonetic law to be formulated on the strength of it; I therefore again apply the method of reference to a parallel and say: In Bat. gaṅit < laṅit, g represents Common IN l as in bugan < Common IN bulan, “moon”.

10. We shall exhibit the Common IN linguistic phenomena from the following points of view: phonetic system, synthesis of sounds into words, word-accent, formal analysis of words, formation of derived words, reduplication of words, and synthesis of words into sentences (i.e. syntax).

11. The second principal fart of the monograph will have for its subject the Original Indonesian language. We style Original IN the fundamental form of speech from which the individual IN languages are descended.

12.   The basis upon which we shall reconstruct the Original IN is the Common IN as delineated in the first part.

13.   Whereas in the first part we shall deal altogether with real facts, the second part will only yield results of hypothetical value.

14.   The contents of this monograph are my own, both as regards substance and method. It is true that here and there in the writings of other scholars I have come across the remark that this or that linguistic phenomenon is to be regarded as being Common IN; but such observations are only to be found sparsely, in no great numbers, and often unsupported by the necessary evidence.

Note.—As it is my wish that my monographs should not exceed a moderate size, I shall not mention everything that I have recognized as being Common IN; but I shall include everything that seems to me specially important.

15.   The following abbreviations will be used:

Bal.	=	Balinese.	Mad.	=	Madurese.
Bat.	=	Batanese.	Mak.	=	Makassar.
Bim.	=	Bimanese.	Mal.	=	Malay.
Bis.	=	Bisaya.	Mkb.	=	Minangkabau.
Bol.	=	Bolongan.	Mlg.	=	Malagasy.
Bont.	=	Bontok.	Pamp.	=	Pampanga.
Bug.	=	Bugis.	Sund.	=	Sundanese.
Day.	=	Dayak.*	Tag.	=	Tagalog.
Form.	=	Formosan.	Tar.	=	Tarakan.
Inv.	=	Inivatan.	Tontb.	=	Tontemboan.
Jav.	=	Javanese.

Note I.—For languages which have short names, such as Bulu, Toba, Karo, and Hova, and also for such as are only rarely cited, as Bolaang-Mongondou, no abbreviations are used.

Note II.—The abbreviations of the titles of periodicals are those used in the Orientalische Bibliographie.

16. The Formosan material is derived from the notices of Happart and Van der Vlis, and from O, Scheerer's "The Batan Dialect"; the Inivatan material, from a MS. vocabulary most kindly presented to me by 0. Scheerer ; the Sumbawarese, from a text published by Jonker in Bijdr. 1904, pp. 273 seqq.

PART I

COMMON INDONESIAN

---

SECTION I: PHONETIC SYSTEM.

17. We have to recognize as belonging to Common IN the six vowels : a, i, u, ĕ, e, and o.

18. The vowel a. We meet with this vowel in the Common IN word kayu, "tree". That kayu is a Common IN word is proved by the following table :

Tree. Philippines, Ibanag: kayu — Celebes, Tonsea: kayu — Borneo, Day.: kayu — Java, Jav.: kayu — Sumatra, Mkb. : kayu — Malay Peninsula, Mal.: kayu — Madagascar, Hova: hazu — Northern Border, Form. : caiou — South-Western Border, Simalurese: ayu-ayu.

Note I. — The spelling in Happart and Van der Vlis' Form, vocabularies is very awkward; we need have no hesitation in interpreting caiou as = kayu.

Note II. — Hova hazu < kayu In accordance with the two following phonetic laws: "Common IN k becomes h in Hova, except after the velar nasal or as a final ". — " Common IN y appears in Hova as z".

Note III. — The disappearance of k in Simalurese ayu-ayu is supported by the parallel case of iuq as compared with the widespread Common IN ikur, " tail ".

Note IV. — The word kayu appears in a number of other IN languages besides the above, e.g. in Sumatra in Běsěmah, Lampong, Karo, and Gayo, as well as in Mkb., and everywhere unchanged in sound. But it will suffice if in each case we select one language out of each of the ten areas of distribution.

19.   We have, therefore, demonstrated the existence of a Common IN word kayu, which contains the vowel a. We also became acquainted in § 1 with a Common IN laṅit, “ sky ”, and later on we shall meet with the Common IN words apuy, “ fire ”, ama, “ father ”, ina, “ mother ”, ratus, “ hundred ”, and a, “ the ”, as well as the Common IN formatives ka-, ta-, and -an. All these forms agree in containing the vowel a. The amount of this material is so large that we may without hesitation pronounce the vowel a to be Common IN.

20.   The vowel i.   This vowel is proved to be Common IN by the word aṅin, “ wind ”, in accordance with the following table:

Wind.   Philippines, Iloko : aṅin — Celebes, Bug.: aṅiṅ — Borneo, Sampit: aṅin — Java, Sund.: aṅin — Sumatra, Toba: aṅin — Malay Peninsula, Mal.: aṅin — Madagascar, Hova : anina — Eastern Border, Bim. : aṅi — South-Western Border, Simalurese: aṅin.

Note I.—Phonetic laws: “ Bug. and Mak. unify all Common IN final nasals into ṅ ”, hence Bug. aṅiṅ < aṅin.—“ Hova unifies all Common IN final nasals into -na ”, hence anina < aṅin.

Note II.—The law in accordance with which Common IN ṅ has become n in Hova, as in anina < aṅin, and the law in accordance with which final consonants disappear in Bim., as in aṅi < aṅin, have already been given. We assume that the reader will make a mental note of all such laws, and we therefore mention each of them only once.

21.   We have become acquainted with a Common IN word aṅin, “ wind ”, which contains the vowel i. In the course of our enquiry we shall meet with many other Common IN words containing the same vowel. But we will content ourselves with specifying this one instance, the word aṅin ; the reader will, of course, notice the other cases; and in the sequel we shall pursue the same method. We therefore pronounce the vowel i to be Common IN.

22.   The vowel u. This is proved to be Common IN by the word kayu, dealt with in § 18.

23.   The vowel ĕ, an indeterminate, rapidly pronounced sound often called by its Jav. name pĕpĕt. This vowel estabhshes its claim to be styled Common IN by the evidence of the word tĕnah, “ half, some ”.

Half.   Celebes, Bug.: tĕṅṅa — Borneo, Bol.: tĕṅah — Java, Sund.: tĕṅah — Sumatra, Karo: tĕṅah — Malay Peninsula, Mal.: tĕṅah — Eastern Border, Sawunese: tĕṅa — South-Western Border, Simalurese: tĕṅah.

Note.—The phonetic laws are: “Common IN final h disappears in Bug., Mak., and several other IN languages ”.—“When in Common IN ĕ, of a non-final syllable, is followed by a single consonant, that consonant appears doubled in Bug.”—Both these laws are exemplified by Bug. tĕṅṅa.

24.   The number of areas in which we have met with ĕ amounts only to seven ; according to the principles enunciated in § 3 we ought therefore to have some hesitation in declaring it to be Common IN.   Whereas a, i, and u are present everywhere, the pĕpĕt is wanting in some IN languages, which instead of it use another vowel: e.g. Day. does not say tĕṅah, but teṅah, Bis. says toṅa. But in several of these languages the pĕpĕt has left tangible traces of its former existence.   In Mak. of Celebes the pĕpĕt appears as a; Mak. therefore possesses two etymologically distinct a’s, the one corresponding to the Common IN a, the other to the Common IN ĕ. But the second a causes certain consonants which follow it to be doubled, the first one does not affect them. So Common IN anu, “ someone ” (§ 135), appears in Mak. as anu, but Common IN ĕnĕm, “ six ”, as annaṅ, and Common IN tĕkĕn, “ staff ”, as takkaṅ. In this matter Bug. exhibits the intermediate stage, for it also doubles the consonant but it retains the pĕpĕt:

Common IN	tĕkĕn
Old and Modern Jav.	tĕkĕn
Bug.	tĕkkĕṅ
Mak.	takkaṅ

In precisely the same way Iloko in the Philippines, which has replaced ĕ by e, doubles the consonant and accordingly says

tekken.—Further, in the Hova of Madagascar the pĕpĕt appears in an accentuated syllable as e but in an unaccented one as i; accordingly Common IN ĕnĕm appears in Hova as enina. Hova therefore has two etymologically distinct i's, the one being Common IN and the other derived from the pĕpĕt. Now before the first of these i’s Common IN l becomes d in Hova, but before the second it remains unchanged. For Common IN lima, "five", Hova has dimi, but for alĕm, "night", alina.

Note.—In the final of Hova dimi we have a third i, originating from a, in accordance with the following law: “ A Common IN final a is assimilated in Hova to an i of the preceding syllable ”. Analogous cases of assimilation have been discussed by Conant in Anthropos 1911, pp. 143 seqq.

25. We have shown in the preceding paragraph that in Iloko and Hova the formerly existing pěpět is still traceable to-day; we therefore add the Philippines and Madagascar to the number of the regions mentioned in § 23, and are thereby entitled to style the pĕpĕt a Common IN feature.^[1]

26. The vowels e and o. In contrast to the pĕpět these vowels have, one might almost say, a universal distribution; though it is true that Hova has no o and Mentaway very few words that contain an e. But e and o have in many cases originated out of other sounds, e.g. in Mad. pote, which stands by the side of Common IN putih, “ white ”. And secondly, words containing an original e or o can as a rule be traced only through a very limited number of languages. Perhaps the most widely distributed are the two words bela, “ companion, avenger, to share the same fate ”, and sor, “ above ” or “ below ”. The e and o in these words are original; at least I know of no indication whatever that they are derived from any other sounds.

To share the same fate. Celebes, Mak. : bela — Java, Jav.: bela — Sumatra, Gayo : bela — Malay Peninsula, Mal. : bela — Eastern Border, Bim. : bela.

Above, below. Philippines, Tag.: anor, "to lift up, to carry" — Celebes, Tontb.: sosor, "to go up" — Java, Old Jav. : sor, "below" — Sumatra, Lampong: ansor, "to diminish" — Northern Border, Form.: masor, "to exceed".
Note I. — Tag. anor with n for s is formed in accordance with the principles illustrated in § 149.
Note II. — The phenomenon that one and the same word means both "above" and "below" finds a parallel in the Rottinese demak, which signifies both "high" and "deep". For other cases see Kern, "Fidjitaal", p. 211.
27. In view of what has been said in the preceding paragraph, it is only with some hesitation that we can venture to style e and o Common IN vowels.
28. We must concede to Common IN the three diphthongs uy, ay, au (which I prefer to write aw).* These appear only as finals. In the interior of words, as in Mal. laut, the two vowels belong to different syllables.
29. The diphthong uy. This is shown to be Common IN by means of the Common IN word apuy, "fire ".
Fire. Philippines, Inv.: apuy — Celebes, West Mori: apuy — Borneo, Tar.: apuy — Java, Old Jav.: apuy — Sumatra, Achinese: apuy — Northern Border, Bat.: apuy — South- western Border, Simalurese: ahoy.
Note.—In Simalurese ahoy we find oy < uy in accordance with the parallel lañoy, "to swim", as compared with Old Jav. lañhuy. — The p has disappeared as in the parallel ulau, "island", beside the very widespread pulaw; the h is to be regarded as the last remnant of the vanishing p.
30. The diphthong uy becomes u in Hova, so we find afu < apuy, walu, "to change, to turn back", beside Old Jav. waluy. But when a suffix is appended to such word-bases as these, the y is no longer a final and therefore need not dis-
*[In Romanized Malay these diphthongs are commonly written ui, ai, and au, respectively. For the reason why it is preferable (at any rate in works like the present) to adopt the author's spelling, see Essay IV, § 158.] appear: it then becomes z in accordance with the law stated in § 18. Now in Old Jav. beside the indicative waluy there is a conjunctive waluya, and with this there corresponds in Hova, according to § 108, the imperative ini-waluza. This formation mi-waluza < mi-waluya therefore proves to us the former existence of the diphthong uy in Madagascar.

31. In view of the evidence set out in §§ 29 and 30 we may pronounce the diphthong uy to be Common IN.

32. The diphthongs ay and aw. About as widely distri- buted as the Common IN apuy are also patay, “dead”, and paraw, “hoarse”. These two words therefore warrant us in regarding the two diphthongs ay and aw as Common IN.

33. The semi-vowels y and w. The former is shown to be Common IN by kayu (§ 18), the latter by means of the word walu, “eight”.

Eight. Philippines, Magindanao : walu — Celebes, Tondano : walu — Borneo, Tar. : walu — Java, Old Jav. : wwalu — Sum-atra, Gayo: waluh — Madagascar, Hova: walu — Northern Border, Form, dialects: walu — Eastern Border, Tettum: walu — South- Western Border, Mentaway: balu.

Note I. — The pronunciation of the w is not uniform every- where; it appears to be chiefly bilabial, the Philippine text-books often write it u or even o, thus ualu, oalo.

Note II. — Phonetic law: “Common IN w appears in Ment-away as b”, hence balu < walu.

Note III. — Old Jav. wwalu < walu in accordance with the parallel wwara, “to exist” , as compared with wara elsewhere (see § 188).

Note IV. — Gayo waluh has got its h through the influence of neighbouring numerals which really possess a genuinely primitive one, such as tujuh, “seven”. Analogous changes in numerals are mentioned in § 141.

34. We must admit as Common IN the relars k, g,ñ.^[2]

35. The velar k is evidenced by the word kayu (§ 18); the velar ṅ by aṅin (§ 20) ; the velar g by the word dagaṅ, “stranger, trader”.

Stranger. Philippines, Tag. : dagaṅ — Celebes, Bolaang- Mongondou: dagaṅ — Borneo, Day.: dagaṅ — Java, Sund.: dagaṅ — Sumatra, Toba: dagaṅ — Malay Peninsula, Mal.: dagan — Eastern Border, Bim. : daga.

36. We must concede to Common IN the palatals c,^[3] j, ñ^[3] But the pronunciation of this series of sounds is not quite identical everywhere : in several langua es they are palatalized dentals, in Bont. the two explosives are " frequently near ds and ts ", as Seidenadel says.

37. The tenuis c. This is shown to be Common IN by the word racun, “poison”.

Poison. Celebes, Bug.: racun — Borneo, Sampit: racun — Java, Jav. : racun — Sumatra, Karo: racun — Malay Penin- sula, Mal. : racun — Eastern Border, Bim. : racu.

38. The media j. This is proved to be Common IN by means of the word jalan, “path”.

Path. Philippines, Bont. : jalan — Celebes, Bareqe : jaya — Borneo, Sampit: jalan — Java, Sund.: jalan — Sumatra, Besemah: jalan — Malay Peninsula, Mal.: jalan — Eastern Border, Sawunese: jara.

Note, I. — Phonetic law: “Bareqe, like Bim. and Sawunese, tolerates no final consonants”, hence jaya < jalan.

Note II. — Parallels: Bareqe jaya with y < l as, in buyu, “mountain”, beside Bug., etc., buluq. — Sawunese jara with r < l as in mara, “exhausted” , beside Mal., etc., malas.

39. The nasal ñ. This is shown to be Common IN by the word añud, “to drift”. To drift. Philippines, Pamp.: añud — Celebes, Mak.: añuq — Borneo, Day. : hañut — Java, Old Jav. : añud — Sum-atra, Besemah : añot — Malay Peninsula, Mal. : hañut.

Note I. — Phonetic laws: “Day., Mal., Besemah and other languages, particularly of Borneo and Sumatra, change Common IN final media into tenuis”, hence Day., IMal. hanut. — “Bug. and Mak. change Common IN explosive final into q”, hence Mak. anuq. — “Besemah renders Common IN ending u -- consonant by o -- consonant”, hence añot.

Note II. — The h in Day. and Mal. hanut is a petrified for- mative: Day., for instance, uses ha-, or h-, to form verbs from word-bases.

40. If we survey once more the area of distribution of the series c, j, ñ, we are compelled to admit that it is not extensive enough to justify us in declaring without hesitation that this series is Common IN. However, we find in Madagascar, in Old Mlg., a further piece of evidence, at any rate for j, in the spelling dz, which we meet with occasionally in Flacourt and may compare with Seidenadel's ds in § 36. Thus Flacourt mentions a word idzin, “dark”  : see Ferrand's edition, p. 103. But this idzin coincides phonetically with Jav. ijem, in accord- ance with the phonetic law: “Common IN and also Old and Modern Jav. final -em becomes -ina in Mlg.; but the older sources often represent it by -in”. See also alina < alem, § 24. — It is true that idzin means “dark” and ijem “green”, but we find in IN more than one parallel for the shift of meaning from “green” to “dark” . In Madagascar itself maitsu signifies “green” in some of the dialects and “black” in others; Mal. hijau means “green” and is also used to describe the glint of black hair. — In the Mlg. of to-day z, i.e. the sonant sibilant, is used for Flacourt's dz, that is, for Common IN j.

Note. — It is remarkable what a number of Day. words con- taining j occur also in Hova, including some that are peculiar to these two languages. Examples : Day. jara, “punishment” = Hova zara, “lot, luck” — jera, " tofrighten away by beating " = zera, “to beat” — jawoh, “negligent” = zawu- zucu, “to behave carelessly” — joho, “wantonness of spirit” = zu “fame” , etc. — I am of opinion that Hova, or Mlg., finds its nearest relative among the IN languages in this very Day. Besides the circumstance I have just mentioned there are a number of other observed facts that have led me to that opinion, such as their common possession of the passive in buah, their peculiar adverbial use of the numerals (§ 172), etc.—Porzezinski's contention with regard to the relationship of Mlg. to other languages (Porzezinski-Boehme, p. 77) is untenable.

41. We must attribute to Common IN the dentals t, d, n. The tenuis is evidenced by laṅit (§ 1), the media by dagaṅ (§ 35) and the nasal by aṅin (§ 20).

42. We must attribute to Common IN the labials p, b, m. The tenuis has been shown to be Common IN by the word apuy (§ 29).

43. The media b. This can be shown to be Common IN by means of the word buṅa, “flower”.

Flower. Philippines, Pamp.: buṅa — Celebes, Mak.: buṅa — Borneo, Bol.: buṅa — Near Java, Mad.: buṅa — Sumatra, Toba: buṅa — Malay Peninsula, Mal.: buṅa — Eastern Border, Bim.: buṅa — South-Western Border, Nias: buṅa.

Note.—Pamp. buṅa does not mean “flower” but “fruit”, and Mad. buṅa means "seed-bud". Compare the parallel case that in Mentaway bua—which, by the way, has no etymological connexion with buṅa—means “flower” as well as “fruit”.

44. The nasal m. This is shown to be Common IN by means of the word ama, “father”.

Father. Philippines, Inv. : ama — Celebes, Bolaang-Mongondou: ama — Borneo, Bol.: tama — Java, Sund.: ama or rama — Sumatra, Gayo: ama — Madagascar, Mlg. dialects: zama — Northern Border, Form. dialects : ama, rama, tama — Eastern Border, Masaretese : ama — South-Western Border, Siberutese: ama.

Note I.—In rama, tama and zama the articles ra, ta and i have coalesced with the word ama: see §§ 93 and 187.—Mlg. zama has undergone the following development : zama < yama, in accordance with § 18, < iama = article i + ama. A parallel thereto is Mlg. zahu, beside ahu, "I", < article i + aku. The intermediate form yaku is preserved in Day.: compare what has been said in § 40.

Note II.—Mlg. zama does not mean “ father " but " uncle ". A parallel thereto is the Gayo ama, which signifies both " father " and " uncle ".

45. We must attribute to Common IN the liquids r and l. The sound l is evidenced by the Common IN laṅit (§ 1); the sound r by karaṅ, " rock, dry ground, reef, coral ". § 190 deals with the pronunciation of the r.

Rock. Philippines, Iloko: kalaṅ — Celebes, Mak.: karai — Borneo, Day.: karaṅ — Java, Sund.: karaṅ — Sumatra, Gayo: karaṅ — Malay Peniusula, Mal.: karaṅ — Madagascar, Hova:harana — South-Western Border, Nias: kara.

Note.—Iloko l for r in accordance with the RLD-law (§ 190). — A phonetic law of Nias: " No final consonant is tolerated in Nias ".

46. We must attribute to Common IN the sibilant s. This is evidenced by the word susu, " breast, to suck ".

Breast. Philippines, Inv.: susu — Celebes, Tontb.: susu — Borneo, Bol.: susu — Java, Sund.: susu — Sumatra, Lampong: susu — Malay Peninsula, Mal.: susu — Northern Border, Bat.: susu — Eastern Border, Tettum: susu — South-Western Border, Nias: susu.

47. We must concede to Common IN the aspirate h. This is evidenced by the word tahan, " to hold fast, to retain ".

To retain. Philippines, Tag.: tahan — Celebes, Ponosakan: mo-tahaṅ — Borneo, Day.: tahan — Java, Sund.: tahan — Sumatra, Lampong: tahan — Malay Peninsula, Mal.: tahan.

Note.—It must be admitted that the distribution of h is of such a nature that we have some hesitation in pronouncing it to be Common IN.

48. We meet with other sounds besides these in the IN languages: thus in § 30 we became acquainted with an f in the Hova word afu, " fire ", and in § 44 with a z in the Hova zahu, " I "; and Bug. and Mak. have the glottal explosive q,^[4] etc.

But on the one hand these sounds are not very widely distributed, and on the other they are demonstrably of a secondary kind. Thus the Hova f originated in accordance with a phonetic law from the Common IN p, as is proved by the comparison of afu with Common IN apuy; and the law is: “ Common IN p becomes f in Mlg., save after a labial or when final ”.   On such sounds as these we cannot confer the title “ Common IN. ”

49.   Common IN, therefore, has the following phonetic system, though it must be admitted that some of the sounds have not been evidenced with absolute certainty:^[5]

a	i	u	ĕ	e	o
k	g	ṅ[6]
c[6]	j	ñ[6]
t	d	n
p	b	m
y	r	l	w
s
h

SECTION II: SYNTHESIS OF SOUNDS
INTO WORDS.

50. Apart from interjections and words of form, the words of the IN languages, as we meet them when we open the dictionaries (i.e. in the shape which is more accurately called the “ word-base ”), are mostly disyllabic. We may say, therefore, that as a general rule the primary synthesis of sounds is into a disyllabic structure. Thus in the Mal. version of the Rāmāyana, where the story is told of how Hanuman was sent to Langkapura, there is the sentence: “ Now Hanuman was sitting under a maja tree ” = Hanuman pun duduq-lah di bawah pohon maja. Here we have four disyllabic word-bases in succession.

Note.—Mal. lah here serves to emphasize the predicate. — di bawah literally means “ at (the) bottom (of) ”.

51. Monosyllabic words of substance, i.e. verbs and nouns, are rare; some languages possess none at all; Old Jav. has the largest number. And we never find one and the same monosyllabic word of substance running through very many different languages. Probably the one that is most widely distributed is kan, “ food ”.

Food. Philippines, Magindanao: kan — Celebes, Tontb.: kan — Sumatra, Pabian dialect of Lampong: kan — Eastern Border, Masaretese : ka — South-Western Border, Mentaway: kan.

Note I.—It must not be imagined that this monosyllabic kan merely figures in the dictionary: it really exists in living speech. We find in the Mentaway story Ägā-mu-la-laibi the phrase: “ There is no food ” = tata kan, and that is a complete sentence. Again, in the Mentaway legend of the origin of the race we read: “ There were plantains for food ” = aiat kan bago; and that, too, is a complete sentence. Note II. — -In other languages kan is found as a radical constituent of disyllabic words of substance, as for instance in the Old Jav. paṅan, “food”, where k has become ṅ in conformity with the principles of § 149. Thus in Jonker's Book of Laws, Art. 15, there is the sentence: “If he is a minor lodging with another person, he is liable for his board ” = “If is child under-age living with a man, owes food” " = yen hana rare alit, aṅherṅher iṅ woṅ, ahutaṅ paṅan.

52. If kan, which only appears in five areas of distribution, is the most widely spread monosyllabic word of substance, we must declare that we are unable to style any single monosyllabic word of substance “Common IN”.

53. There are also trisyllabic word-bases, but these, too, are not numerous, and we seldom find any one of them running through a number of languages. Perhaps the most widely distributed one is the word banua, “district, inhabited place”.

District. Philiphines, Pamp.: banua — Celebes, Bareqe: banuwa — Java, Old Jav. : wanwa — Sumatra, Toba : banua — Malay Peninsula, Mal.: benuwa — South- Western Border, Nias: banua.

Note I. — Old Jav. -w for b as in the parallel wuṅa, “flower”, for Common IN buṅa (§ 43).

Note II. — Between the u and the a some of the languages have developed the transitional sound w, hence Bareqe banuwa ; in Old Jav. the u before the a has become a consonant, hence wanwa. An exact parallel hereto is afforded by Common IN buah, “fruit”:

Without transitional sound : Nias banua, bua.

With transitional sound: Mai. bĕnuwa, buwah.

U turned consonant: Old Jav. wanwa, wwah.

Note III. — In Mal. a full vowel preceding the accent is weakened into the pepet. Parallel : the loan-word sĕrdadu < soldado, “soldier”.

Note IV. — Pamp. banua means “sky”. We have a parallel to this transference of meaning in Bis., wherein banoa denotes both a region of the earth and a region of the sky. Li ice wise in Toba banua ginjan = " Upper banua " = " sky ", while banua tona on = " this middle banua " == " the earth ".

54. The initial of words. Every sound of the phonetic scheme set out in § 49 can serve as the initial of a Common IN word. Evidence in support of this is superfluous. But there cannot be more than one consonant at the beginning of a word. Initials of two consonants appear in quite isolated cases; thus Gayo has contracted Common IN beli, " to buy ", into bli*

55. The interior of words, or more precisely, the conson-antal element between the two vowels of disyllabic words. In this position every individual IN language admits one or two consonants, but not more.

56. The commonest case of the combination of two con-sonants is that of a nasal + a cognate explosive, e.g., n + k in Tonsea dunkud, "to speak ironically". This case occurs in every individual IN language, and can straightway be styled Common IN.

57. Another common case of the combination of two con- sonants in the interior of a word arises from the reduplication of the root, which is one of the methods of forming words out of roots. Thus in Old Jav. from the root kab, " to move to and fro ", which does not occur in actual speech, we find the derivative formations unkab and, with reduplication, kabkab. Here we have the combination b + k.

58. Now some of the IN languages only tolerate the first of the two above-mentioned cases : nasal + cognate explosive. Others also admit the second one, e.g. certain languages of the Philippines, Java, Sumatra, and the Northern and South- western Borders. Thus Bis. in the Philippines says kapkap for " to touch ", but Day. in Borneo says kakap, and cannot say otherwise.

59. But we find certain indications which render it prob-able that the languages with the kakap type of combination did in a previous stage of their development possess the kap-kap type instead of it.

The most important of these indications is based on such phenomena as are set out in the following table:

Mad.	taptap	Mad.	kapkap
Mad. dialect	tattap	Mad. dialect	kakkap
Sund.	tatap	Day.	kakap

Note.—taptap means “to strike with the flat of the hand”, and the like; kapkap, “to scratch, to touch”, and the like.

We have, therefore, in particular dialects of Mad. a transitional form, produced by assimilation, and accordingly think it credible that the languages which now only possess the kakap type have evolved it out of a pre-existing kapkap type.*

60.   The result of the considerations in §§ 55-59 is: —Common IN tolerates one or two consonants in the interior of a word; in the latter case we find, on the one hand, the combination of nasal + cognate explosive, and on the other the kapkap type of combination.

61.   The final of words. In Common IN any of the vowels may serve as a final; evidence in support of this is superfluous. Secondly, the diphthongs uy, ay and aw, as was shown in § 28 seqq. The investigation of the consonantal finals is a more complicated matter.†

62.   From this investigation we must first exclude the palatals. The consonants of that series are incapable of doing duty as finals in any IN language, and accordingly it must be declared that Common IN does not tolerate final palatals. — Tontb. has a few words with final c, e.g., paliqpic, " a certain part of the roof ". But this c is a secondary formation originating from k in accordance with the Tontb. law formulated by the two Adriani's: " In Tontb., k after i becomes c ". The original form with k, palikpik, occurs in Tonsea, a language closely related to Tontb. Note.—The first k of palikpik appears in Tontb. as q in accordance with the following law, also formulated by the two Adriani's: "In the case of a Common IN combination of consonants, other than nasal ＋ cognate explosive, the first of the two consonants becomes q in Tontb."

63. Setting the palatals aside, we find in the individual IN languages very various possibilities of consonantal endings, which the following three typical extracts from texts will at once serve to illustrate:

I. Sĕraway sentence, out of the Anday-Anday Riṅgan Sĕdayu: "The king had a son and a daughter" = King had son had daughter = rajaw bĕranaq bujaṅ bĕranaq gadis.

II. Nias sentence, out of the heroic song edited by Lagemann: "He is fallen into the broad sea" = Finished fallen lie into sea broad = no aeχu* ia ba nasi seholo.

III. Mak. sentence, out of the Jayalangkara, p. 101: " He did obeisance, bowing his head down to the ground " = He did obeisance, bowing head his down to earth the = na aqñomba, sujuq ulu-nna nauṅ ri butta ya.

In the first sentence all the words end in consonants, even the loan-word raja, which elsewhere always terminates in a vowel, has acquired a final w in conformity with the phonetic law: "Final a of other languages appears in Sĕraway as aw".† In the second sentence none of the words ends in a consonant. In the third sentence consonantal endings are in the minority.

Now among the various individual IN languages we notice three principal types of consonantal ending: Some languages tolerate no final consonants at all, or only very few, Bug. for instance, two, viz. q and ṅ; other languages admit all the consonants, except the mediæ; others, again, allow all the consonants, including the mediæ, to serve as finals.

64. The first group, which tolerates no consonants, or very few of them, as final sounds — Bug., Bim., Nias, Hova, etc., languages which, by the way, are not closely related together — does not represent the Common IN condition. For it can frequently be proved that in these languages final consonants have become mute, i.e. they existed in a former stage of the evolution of the language; this can be shown by the evidence of derivative words built up from word-bases by means of sufiixes. Common IN nipis, "thin", appears in Hova as nifi, having lost its s. From nifi is formed a verb manifi, "to make thin", and this forms its imperative with the suffix -a, as in § 30, but that imperative is not ma-nifi-a: it is ma-nifis-a, because here the s, having shifted into the interior of the word, is no longer liable to be affected by the laws that govern final consonants. Here, then, we have evidence that Hova also originally said nifis < nipis. And cases of this kind can be adduced in considerable numbers. But I have failed to discover in these languages any evidence of the former presence of final mediæ.

Note. — Progressive restriction in the choice of consonants serving as finals in the case of an Austroasiatic language has been illustrated by Blagden in JA, 1910, p. 498.

65. The second group includes languages which tolerate as finals all the consonants, with the exception of the mediæ. Where the languages of the third group exhibit mediæ, those of the second have tenues; thus Bis. bokid, "hill", is represented by Mal. bukit, and lawod, "sea", by laut. To this group belong in particular certain languages of Borneo, Sumatra, and the Malay Peninsula, languages, therefore, whose territories lie near to one another, and, furthermore, languages which are at any rate in part somewhat closely connected together. Nevertheless, in contrast with the languages of the first group, we are here in a position to show that these languages of the second group originally also possessed final mediæ. Only we cannot for this purpose use the evidence which served us in dealing with the first group, namely the extension of disyllabic word-bases by means of suffixes. Even when the above-mentioned word laut, " sea ", receives a suffix, e.g., -an, the tenuis remains and no media appears, the result is lautan, "ocean", and similarly in all the other cases. The evidence to which we must now have recourse is furnished by the formative process that has created disyllabic word-bases out of monosyllabic roots. We find in Old Jav. a root rug, which is used by itself without any further extension (i.e. after the fashion of kan in § 51) as a word, with the meaning of " devastated ". Mal. also possesses this root, though not as an independent word, but only as a con-stituent embodied in disyllabic word-bases. When it is com- bined with a prefixed element, the media g is a final, and must therefore change into the tenuis; but if the root is linked with a sufiixed element, the media appears in the interior of the word and is preserved. Thus we get, on the one hand, Mal. buruk < bu + rug, " to fall to bits ", and on the other, rugi < rug + i, "to damage".^[7]—Other Mal. examples: sigi, "to dig with the fingers", beside Old Jav. sisig, "to rub with the fingers", tubi, "to persevere in a thing", beside the Old Jav. tub, which has the same meaning.—These word-bases rugi, sigi and tubi, therefore, tell us that the Mal. of the Malay Peninsula in an earlier phase of its development tolerated the mediae at the end of words.

Note.—The element -i, which occurs in rugi, is a very common phenomenon; it serves both to form word-bases out of roots, and also to give a further extension to word-bases (see § 156). The element bu- only serves to make word- bases out of roots, and it is of rarer occurrence; therefore we will add a parallel: From the root way, "to rock to and fro", Old Jav. forms a-way, "to wave", and Tag. has bu-way, "to see-saw" .

66. The same kind of evidence as has been given for Mal. can also be produced in the case of certain languages of Sum-atra, e.g. Karo, Toba and Mkb.' Old Jav., for instance, has a word-base anteg, "to arrive at", with a root teg; Toba has togi < tog + i, “to conduct to a place”, with, o for e in con-formity with the law: “Common IN pepet appears in Toba as o”.— For Borneo, too, we have similar evidence: we merely add that rugi also exists in Day.

Note. — It will be seen in § 156 that the formative element -i serves to make transitive verbs, and it accords with this principle that the word togi is transitive, whereas anteg is intransitive.

67. The conclusion to be drawn from the facts set out in §§ 65 and 66 is: The languages of Borneo, Sumatra and the Malay Peninsula in an earlier stage of their development also used the mediae as finals.

68. The third group comprises the languages which tolerate all the consonants as finals, including the mediae. To this group belong languages of the Philippines, Celebes, Java, and the Northern and South-Western Borders. But even in these the use of the mediae as finals is not of very frequent occur-rence, so we shall not rest content with asking the reader to glance at the dictionaries, but will give a few actual details of the matter.

69. We meet with final medias more frequently in Old Jav. and the Philippine languages than elsewhere. Ex-amples :

“To manage, to take pains over” " : Old Jav.kĕpug, Bis. kopog.

“Model, pattern” " : Old Jav. tulad, Pamp. tulad.

“To conceal” : Old Jav. kubkub, Pamp. kubkub.

Note. — Phonetic law: “Common IN ě and Common IN u become o in Bis.” , hence kopog.

70. Examples from the languages of Celebes:

Ponosakan bowog < bobog, beside Jav. bog, “to strike”.

Tonsea tuud, beside Old Jav. tuwed, " “stump of a tree”.

Tonsawang kokob, beside Old Jav. kubkub, “to conceal” .

Note. — For the correspondence of u and wĕ in tuud: tuwĕd we have a parallel in Mal. laut: Karo lawĕt, " sea ". It must be admitted, however, that the parallel is not very con-clusive, inasmuch as it does not occur in the same two languages^[8]

71. Examples from the Northern Border. The Form, dialects are rich in cases of final mediae, but there is often a difficulty in finding parallels for them in other IN languages and thus correlating them with IN. The cases for which no such analogues can be discovered might, after all, be loan-words from non-IN forms of speech. Therefore it has seemed advisable to give a somewhat longer list in this connexion:

Form, dobdob : Tag. dobdob, "to poke the fire".

Form, laub : Bis. laob, "to roast".

Form, soab : Old Jav. suwab, "to yawn".

Form, abad : Pamp. babad, "to become damp".

Form, utod : Bis. otod, "crippled".

72. Examples from the South-Western Border. The Mentaway dictionary registers more than a dozen words with final mediæ, e.g., jud-jud, "high water", which is etymo-logically related to the añud of § 39, and bäb, "to hit (the goal)", which is identical with the Achinese bĕb, "to fall upon" (ef. Snouck Hurgronje, "Studien", p. 62).

73. To sum up the results of this discussion on the final mediae (§§ 65-72), we have succeeded in showing that in the Philippines, Celebes, Borneo, Java, Sumatra, the Malay Pen-insula, and on the Northern and South-Western Borders, that is in eight of our regions, there are languages which admit all sounds as finals, including the mediæ, or else formerly admitted them. Hence the conclusion: In Common IN any sound can be a final, always excepting the palatal consonants.^[9]

74. Let us now just recapitulate the propositions estab-lished in the present Section, "Synthesis of Sounds into Words"; The Common IN word, apart from interjections

and words of form, is usually disyllabic. It may begin with any sound in the Common IN phonetic system. In the interior of it, between the two vowels, there may be one consonant, or two consonants, and in the latter case we find nasal + cognate explosive or else the manifold combinations of the kapkap type. At the end any sound can occur, excepting the palatals.

---

75. The great majority of the IN languages accentuate their words, whether they be word-bases or extensions there-'of, on the penultimate syllable. This must be regarded as the Common IN condition.*

76. Accentuation of the final syllable is found in three cases in the various individual IN languages:

77. The first case : The languages which still possess the pepet usually accentuate the final syllable when the pen-ultimate contains a pĕpĕt, e.g., lĕpas, “free”.Ṫ

78. The second case : When a monosyllabic word-base is extended by means of a prefix, the accent in many IN lan-guages remains on the final syllable, i.e. on the word-base. ‡

Accentuation of the word-base. Celebes, Bug.: maqnoq, “to descend” <; formative maq + word-base noq — Java, Jav.: uwos, “rice” < u + wos, see Poensen, Jav. Gr., p. iZ —Sumatra, Toba: mandok, "to speak", word-base dok — Mada-gascar, Hova: wuala, “denied” , word-base la — Eastern Bor-der, Bim.: kambe, “to bleat”, word-base mbe — South- Western Border, Mentaway: patok, “to draw” , word-base tok.

79. The third case : In many IN languages the vocative is accentuated on the final syllable.

Vocative accentuation. Sangir group, Sangirese: aman, “o father!” beside aman, “father” — Celebes, Gorontalese: nana, “o mother !” — Near Java, Mad. : pateq, “thou dog !” — Sumatra, Toba: aman, “o father!” — Eastern Border, Bim.: ind, “o mother I” — South- Western Border, Nias: ind, “o mother !”

Note I. — Example of a sentence with vocative accentuation, from Breukink's Gorontalese dialogues : “Mother, come here I” = nanà poolo.

^[10]

Note II. —It must be admitted that the second and third cases of the accentuation of the final syllable are not so widely distributed as to entitle us unhesitatingly to style them Common IN. On the other hand, it must be borne in mind that in many grammars this very question of accentuation has received the most inadequate treatment. Thus Harde-land says: “The accent remains on the final syllable of such few monosyllabic words as there are, even when they are extended into disyllables or polysyllables by means of pre-fixes”, and this would suffice to establish the case of the accentuation of the final for Borneo as well; but the instances he proceeds to give are dubious: hàì, whence kahàì, and the like must surely be disyllabic forms. — Again, Kruyt's Bareqe Grammar says nothing about the accentuation of the voca-tive, yet Adriani in Ts. Lid. 1. 1. vk., 1910, p. 211, quotes voca- tives of that kind, e.g., oñgà, " friend ! " (addressed to a woman).

80. Words of form often lean proclitically or enclitically on the words which they accompany, and hence they are often written continuously with them in texts in the native alphabets. Thus we read in the Mak. epic Maqdi, towards the end: “And the burial service was read over his head” = na nibaca mo talakkin a ri ulu-nna. The words of form are: na, “and”, mo, emphatic particle, a, article, ri, “over” , nna, “his”. But in the original (Mak. Chrestomathy, p. 426) this sentence is written in three “complexes” or conglomerations, viz. {n + n + i + b + c + m + o) —{l + l + k + i + n) — (r + i + u + l + u + n).^[11] Hence, also, it is not uncommon for words of form to coalesce with the words they accompany, as illustrated by rama, tama and zama in § 44, and other cases which we shall meet with later on. And it is in this way that words of form have in some instances be-come formatives: the passive formative ka- is really the pre-position ka, and Bug. kacalla, “to be accursed” ,properly means “(to come) into a curse”.^[12]

Preliminary Observations.

81. Regarded from the point of view of their formal structure, the words of the IN languages fall into five classes: interjections, words of form, words of substance,pronouns and numerals.

Interjections.

82. The interjections found in the several IN languages are mostly monosyllabic formations, incapable of being ana-lysed further. They can end either in a vowel or a consonant. Example: the Tontb. Dirge for a Dead Mother begins: “Alas, mother, o mother, o mother !” " = o inaq, e inaq, e inaq. They often appear in reduphcated form, e.g. Sund. bobo beside bo,. a word used in mild reproof; and the reduplication may be accompanied by vowel change, particularly in cases where the word is intended to imitate a sound (onomatopoeia). e.g. Day. pikpak beside pak, “smack ! pop !"”

83. We will take a closer view of one single interjection, one that we are in a position to style Common IN. It is the interjection of affirmation, and its form is a, or when redupli-cated aa, or ia.

Yes. Philippines, Iloko : a — • Borneo, Day.: ia — Near Java, Bal. : a — Sumatra, Gayo: a — Malay Peninsula, Mal.: iya — Madagascar, Tangkaranese : ia — Eastern Border,. Kamberese: a or aa.

Words of Form.

84. Like the interjections, the words of form are mostly monosyllables incapable of further analysis. They often consist of two sounds, and in that case usually end with the vowel. This characteristic is illustrated, for example, by the words of form contained in the following sentence from the

Old Jav. Śakuntalā : “ He felt as if the mainstay of liis heart were being cut off ” = It was as if was being cut off now stalk of heart his = kadi hiniris ta nāla ni hati nira.

Note.—kadi, “ it was as if ”.—hiniris, passive of hiris, “ to cut ”.—The use of nāla in this sense may be compared with the Malay taṅkay (hati).

In contrast with the interjections, the words of form are seldom reduplicated; one of the rare cases being the doubling of the negative ta to form tata in Mentaway. On the other hand they are capable of entering into the most manifold combinations with one another. Thus the two articles i and tu occur combined in Hova as itu, in Taimuruna as tui, and in Tontb. as iitu. These composite articles do duty as demonstrative pronouns.—Amongst the words of form we shall consider the articles, the prepositions, the a dubitativum, and the negative.

85.   The article i.   Philippines, Tiruray: fantad, “ earth ”, i fantad, “ the earth ” — Celebes, Bug.: i Diyo, “ Madam Diyo ” * — Borneo, Tar.: i amaq, “ the father ” — Java, Old Jav.: i bapa, “ the father ” — Sumatra, Toba: pidoh i, “ the bird ” — Madagascar, Hova: i Butu, “ the (young man) Butu ” ^[13] — Eastern Border, Kamberese: i ama, “ the father ” — South-Western Border, Mentaway: ka i tuan, “ to the master ”.

Note I.—The position of the article in Toba, in pidoṅ i, finds its parallel in another article, viz. the article e, which in Bug. is put after the noun, whereas in Nabaloi it precedes “ The house ” is bola e in Bug., but e baley in Nabaloi.

Note II.—The i in Toba is more accentuated, and is therefore a demonstrative, but when we look through a Toba text, such as the Mula ni debata idup in Meerwaldt, we see that it occurs extremely frequently, and does, after all, perform the functions of an article.

86.   As already remarked in § 84, the article i also occurs as a component part of the demonstrative pronoun itu, and

this is also Mal.; it is likewise a constituent of the demonstrative ai, and this is Form. So Ave may now add the Malay Peninsula and Formosa [i.e. the Northern Border) to the eight areas of distribution set out in § 85. We therefore pronounce the article i to be Common IN.

Note. — Specimen sentence with the article i : Tarakan, from the Story of the Tailed Man: "He ordered his wife to go to Silimbatu " = It was ordered the wife his to go to Silimbatu = sinusub i andu na makaw da Silimbatu

87. The article a. Philippines, Ibanag : tolay a mapia, " man a good " = " a good man " — Celebes, Mak.: jaraṅ a, "the horse" — Sumatra, Gayo: anak bujaṅ a, "boy big the " = " the youth " — Northern Border, Form. : kuiri a rima, " left the hand " = " the left hand " — Eastern Border, Rottinese: nau a, " the grass ".

88. The article a is also a component of the Old Jav. article aṅ < a + ṅ, which is used pretty interchangeably with the simple ṅ: e.g., aṅ anak or ṅ anak, "the child". Similarly it is a component of the pronoun anu, " somebody "(§ 135), which occurs in nearly all the IN languages.

89. Like the article i (§ 44), the article a often becomes indissolubly attached to substantives. Thus beside the Old Jav. buṅ, " sprig ", there is the Common IN buṅa, " sprig, flower, fruit " (§ 43); beside Old Jav. luh, " tear ", the Bagobo luha. A particularly characteristic case is that of pus, " cat ", in Mad., which in that language serves only as a vocative, whereas puts in Day. does duty in all syntactical relations.

90. From what has been said in §§ 87-89 it follows that we must attribute the article a to Common IN.

91. The article ra. This occurs as a living element of speech in a few languages only: Java, Old Jav.: ra Hyaṅ " the deity " — Madagascar, Hova : ra Be, " Mr. Big ".^[14]

92. The article ra (as was first shown by Kern, "Fidjitaal ", p. 163) has coalesced with the Common IN word tu, "master, lord", to form ratu, and this word ratu occurs in very many IN languages.

93. Further, the article ra is found with particular frequency in inseparable combination with words of relationship, especially ama and ina, wherewith it forms rama and rena, which have been discussed in particular by Kern on several occasions. The word rena is found in Java, in Old Jav. ; in Madagascar, in Hova, under the form reni ; and on the Northern Border, in Form.

Note. — Hova ray, "father", is ra + ayah, reni, "mother", is ra + ina, raha, "brother", is ra + aka. Ina is Common IN, ayah is Simalurese, etc., and aka, "brother", Sund., etc. — The phonetic characteristics of reni are explained in § 24 ad fin., those of raha follow from the law given in § 18, and those of ray from the following law : "Common IN a + y + a, or a + y + a + consonant, appear in Hova as ay". Another instance is the Hova word lay < Common IN layar, "sail". This law involves a limitation of the y-law of § 18.

94. From what has been said in §§ 91-93 it is plain that we may style the article ra a Common IN word.

95. The article i is in most languages a personal article, and as such it precedes personal names, words of relationship, and personal pronouns; the article a usually accompanies names of things; the article ra is an honorific particle. — These characteristics may be regarded as being Common IN.

96. The prepositions. Preliminary observation. Article and preposition are in a certain measure identical in IN. That which in one language is an article serves in another as a preposition; thus the two articles i and a, which we have just discussed, are prepositions in Hova, which language possesses the article ni: e.g., a luha, "in front", i masu, "before (the) eyes". Indeed, even in one and the same language it may happen that a word is both an article and a preposition: e.g. Bug. i Diyo, "Madam Diyo", i liwěṅ, "on the other side”. — From §§ 86 and 97 it follows that we must

recognize in the double function of ἰ, as both article and pre- position, a Common IN feature; but in the case of rɑ, on the other hand, only one function appears to be Common IN, viz. its use as an article.

The articles can also serve as unemphatic pronouns of the third person: e.g. Bug. soroq ἰ, “he recedes”; Mak., from the Zamenspraken (dialogues), p. 35: lino i, “it is calm (weather)”. Note. —If the article in Mlg. is ni, why does it not appear in the above-mentioned a luha and i masu? The answer is: In phrases of the nature of formulas the article can be omitted.

97. The 'preposition i. Philippines, Tag.: i baba, “at the bottom” — Celebes, Bug.: i liwěṅ, “on the (other) side ” — Borneo, Day.: i wa, “at the bottom” — Java, Old Jav.: i sira, “by him ” — Sumatra, Gayo: i Gayo, “in the Gayo country ” — Madagascar, Hova: i masu, “before the eyes ”. Note.— In Tag. and Bug. the preposition i is no longer a really living element of speech, it is only found in certain formulas; but from the point of view of our monograph that is immaterial.

98. Kamberese possesses a preposition la, “at, by ”, with a secondary form lai. This lai is the preposition la + the preposition i. — Herewith we have a piece of evidence for the existence of i on the Eastern Border also. Note.—That we have rightly explained lai by la + i, is proved by the following parallel: Old Jav. combines its two prepositions i and ri into iri, and all three forms have pretty much the same meaning, viz. “at, to ”. See also Hazeu, Gayo Vocabulary, p. 532.

99. The preposition i also does duty as a word-base for verbal forms. The one we meet with most frequently is ma-i, “ to go (to) ”, with the formative ma-, for which see § 148. This verb is found in many languages, from Form, on the Northern Border to Simalurese on the South- Western one. So now we have evidence for i in two more areas of distribution. Note. — That our explanation of the verb mai is the right one is proved by the parallel matu < ma + preposition tu, "to". This verb signifies "to set about (doing a thing)"; in Bug. it indicates the future.

100. The facts set out in §§ 97-99 justify us in pronouncing the preposition i to be Common IN.

101. Specimen sentence with the preposition i: Old Jav., Sang hyang Kamahāyānikan, a 49: " The space between the lower and the upper row of teeth " = Space of the teeth, in the lower part, in the upper part = sela ni ṅ huntu i sor i rukur.

102. The preposition n. Philippines, Tag.: an tawo, " the man ", n an tawo, " of the man " — Near Celebes, Talautese, the Cursing of the Fowl, third sentence from the end: laia n awaqa, " heat of the body " — Borneo, Day. : bau n andaw, " face of the sky ", i.e. " cloud " — Java, Old Jav.: tanaya n tani, "people of the district" — Sumatra, Gayo: gêral n guru, "name of the teacher" — Madagascar, Hova: ra n usi, " blood of goats " — Northern Border, Bat. : chinamaṅ-anak^[15] n i santa Maria, " born of the blessed Mary " — South Western Border, Mentaway : uma n abak, " house for boats ".

On the strength of this evidence we may pronounce the preposition a to be Common IN.

103. The preposition ka. Philippines, Bagobo: ka kuda," to the horse " — Borneo, Day. : ka Sampit, " to (the place called) Sampit" — Java, Sund.: ka Banduṅ, " to (the place called) Bandung" — Sumatra, Toba: ha duru, "on to the side " — Malay Peninsula, Mal. : ka darat, " to the mainland " — Madagascar, Hova: ha tratra, " up to the breast " — South-Western Border, Mentaway: ka lagay, " into the village ".

Note. — -Phonetic law: " Toba changes Common IN k into h, save after a nasal or when final, in which cases k persists ", hence ha < Common IN ka. 104. The preposition ha is also used as a conjunction meaning "in order to, until"; Mkb. example: "Rice to be eaten " = nasi ka dimakan. In some other languages a verb maka has been formed from it, after the fashion of mai in § 99, with the meaning "to have an object, or a task". Mak. example: " Has he the task of reading it ? " = Has task he, to read it ? = maka iya lambaca i. Thereby ka is shown to exist in Celebes also; and see § 80 for additional evidence.

105. From the facts set out in §§ 103 and 104 we acquire the right to pronounce ka to be a Common IN preposition. 106. As is shown by the examples in §§ 97-104, the preposition i as a rule indicates the place where, the preposition n the place whence, the genitive, and the preposition ka the place whither. This usage may be regarded as Common IN.

107. The a dubitativum, a word of form used conjunctitively, dubitatively, interrogatively, to weaken the force of a proposition, comparatively, and disjunctively. The a dubitativum, Philippines, Bis.: "What manner of snake is this ? " = Which snake what this ? = onsań halas a kana — Celebes, Bug. : " Is this a slave or a freedman ? " = Slave + interrogative particle g for ga + this or freedman = ata-g-iro a maradeka — Java, Old Jav. : " Like a tongue " = Tongue as = ilat a — Sumatra, Mkb.: " Whatever it may be " = Something, whatever = barań a — Eastern Border, Kamberese: " Only a little " = hakudu a.

108. Besides the dubitative there is also an imperative a: Northern Border, Form.: madis-a, " Hasten ! ", from the word-base madis, " quick " — Madagascar, Hova: mi-waluza, discussed in § 30. We may assume that the imperative use has grown out of the dubitative one delineated in the preceding paragraph. We are justified in doing so when we note the fact that whereas in Old Jav. a, besides indicating comparison, also forms the conjunctive, in Modern Jav. it forms the conjunctive and the imperative, and in Hova the imperative alone. — If we now add the imperative a of the Northern Border and Madagascar to the dubitative a of the preceding paragraph, we may pronounce this particle to be Common IN.

109. Specimen sentence with an a dubitativum, used conjmictively : Old Jav., Mahābhārata, Āśramawasanaparwa, 13 : "It is not seemly that thou shouldst come with us" = Not seemly, thou shouldst-come- with with us = tan yogya kita milwa ri kami. — The indicative is milu, the conjunctive milwa < milu + a. — An a dubitativum, used to limit the force of the verb: Kamberese, from the Story of the Civetcat: "Come here, we wish to deliberate a little !" = Come thou, we deliberate a little only! = mai kaw, ta bataṅ hakudu a.

110. The negative. Among the negatives of the several IN languages, di, either standing by itself or used as the nucleus of a word, has the widest distribution; we therefore style di Common IN.

The negative di, "not". Philippines, Tag.: di — Celebes, Bolaang-Mongondou: diya — Borneo, Day. : dia — Near Java, Mad.: ĕnjaq — Sumatra, Sĕraway: ĕndiaq — Malay Peninsula, Mai.: jaṅan — Madagascar, Hova : dia-hue, "not so".

111. Here we have a Common IN nucleus di, which is mostly accompanied by a formative a, hence Day. dia, Bont. adi. For this attendant a we have many parallels in IN. In several languages there is also a negative ti (which, by the way, does not result by phonetic law from di), as well as another negative ta; and beside these short forms we find in Bulu a form tiya < ti + a, and in Mentaway a form ata < a + ta.

This attendant a is the article a; that appears from the following parallel: The Bug. negative is deq, but the Wajorese dialect of Bug. says deq-sa, and sa is a weak demonstrative in Bug.

112. In Mai. janan < di + aṅan and in Mad. ĕnjaq < ĕn + di + aq, as the Sĕraway still says, the i before the vowel a was first weakened into the consonantal y as in Day. yaku < i + aku (§ 44), and then d + y became j. — How the q in Sĕraway ĕndiaq and Mad. ĕnjaq, beside Day. dia, is to be interpreted, I cannot say: I can only point to the parallel fact that Bug. ajaq, "lest", also has a q, while Old Jav. aja has not^[16]

113. Specimen sentence with a negative: Bont., Kolling, near the end: "Come thou down, that we may eat! Then came he not" = Down, thou, that eat we; then not = banad ka ta maṅan tako; isaed adi.

114. Among the conjunctions we can hardly discover a case that we may venture to call Common IN. Though the conjunction pa is very widely distributed, it has such very different meanings in the several languages that the matter becomes quite uncertain.

Words of Substance.

115. Words of substance — verbs, substantives and adjectives — are mostly disyllabic. They contain a monosyllabic material nucleus, which we call the root, and a formative element ; or else they are formed by the reduphcation of the root, or by the union of two different roots. It seldom happens that the monosyllabic root by itself does duty as a word of substance, like kan in § 51.

116. We will now, in the first place, take an individual joot and show that it is Common IN, choosing for that purpose the root suh, "to enter, to force oneself into, to strike into", and the like.

Root suk. Phihppines, Pamp.: tusuk, "to pierce through" — Celebes, Mak. : usuq < usuk, "to pierce with a needle" — Borneo, Day.: masok, "to enter, to become". — Java, Old Jav. : asuk, "to bring into" — Sumatra, Karo: pasuk, "to knock in" — Malay Peninsula, Mal.: masuq, "to enter, to take sides with a party, to be on a person's side" — Madagascar, various dialects: isuka, "to become engaged".^[17]

117. In accordance with the method dehneated in the Introduction, we may pronounce the following roots, amongst others, to be Common IN : kan, root for word-bases signifying “ food, to eat, to give food to ” — kit, “ pain, to pain, to punish ” — kis, “ to file ” — nis, “ to howl ” — tuk, “ to knock ” — tuṅ, “ to hang ” — tut, “ flatulence ” — num, “ to drink ” — pas, “ free, loose ” — buk, “ dry rot, worm that burrows in wood ” — bah, “ to revere, to pray to ” — raw, “ sun, day ” — run, “ beak, nose, handle ” — laṅ, “ to wind, to twist ” — lik, “ to turn back ” — lit, “ skin, to peel ” — lĕm, “ to dive, evening ” — sih, “ pity, love, to love ”. — For the concept “ to live ” there is no Common IN root.

118. It may happen that only the root itself runs through a number of different languages, while the elements that accompany it vary, as in the case of the root suk in § 116. Or both parts, the root and the formative element, may extend through a number of languages. From the root lit mentioned in § 117, there is formed the verb salit, “ to peel ”, which only occurs in a few languages; but, on the other hand, the same root also goes to form kulit, “ skin ” (and also “ to peel ”), which we must class as Common IN.*

Skin. Phihppines, Iloko: kulit — Celebes, Tontb.: kulit — Borneo, Sampit: kulit — Java, Sund.: kulit — Sumatra, Lampong: kulik — Malay Peninsula, Mal.: kulit — Madagascar, Hova: huditra — Eastern Border, Tettum: kulit — South-Western Border, Mentaway: kulit.

Note.—Lampong kulik beside Common IN kulit in conformity with the parallel: laṅik beside Common IN laṅit.

119. In accordance with our method, the following words of substance, amongst others, can be shown to be Common IN, in their complete disyllabic form, sound for sound: laṅit, “ sky ”, bulan, “ moon ”; but not “ sun ” — apuy, “ fire ”, tunu, “ to burn ”, aṅin, “ wind ”; but not “ warm ” or “ cold ” — buluh, “ bamboo ”; but not “ plant ” — lintah, “ leech ”; but not “ animal ” — ulu, “ head ”, mata, “ eye ”, kulit, “ skin ”; but not “ foot ” — ina, “ mother ”, ama, “ father ”, anak, “ child ”; but not “ step- ” (mother, etc.) — takut, “ fear ”; but not “ joy ” — pilih, “ to choose ”; but not “ to wish ” — tĕnun, “ to weave ”; but not “ to spin ” — tĕkĕn "staff", tali, "cord", suliṅ, "fife" ; but not "hammer" — putih, "white"; but not "red" — tĕṅah, "half"; but not "whole".

120. The following Karo sentence from Si Laga Man: "Then they saw the half of the stone dug out, and now they apphed the lifting pole to it" = ĕṅgo me si tĕnah batu idah ikuruk, e maka ioṅkil na: contains five words of substance, of which two, tĕṅah, "half", and batu, "stone", are Common IN, and precisely in that identical form; but the other three, idah, "to see", kuruk, "to dig ", oṅkil, "to apply a lifting pole", are not.

Pronouns.*

121. The pronomis are very often disyllabic; they are mostly combinations of a specifically pronominal nucleus with formatives, which are mostly articles. Thus, as Seiden-adel has shown, the Bont. pronoun sika, "thou", consists of the article si and the nucleus ka.

122. The monosyllabic nuclei also have an independent existence, in one language or another, but as a rule they aie not very widely distributed. The disyllabic anu (§ 135) is Common IN, while the monosyllabic nu is found in a few languages only, e.g. in Sund. The monosyllabic forms of the personal pronouns have recently been discussed in ex-emplary fashion by Jonker.

123. Specimen sentences with long and short forms. Bug. letter Nomoroq 13, in Matthes: "I have nothing of the kind at home" = But not anything thus I have = nae deqsa anu maqkuwa u taro. — Sund., Nyai Sumur Bandung, p. 66: "We will tell now of her who dwells in Bitung Wulung" = It is told, who dwells in place Bitung Wulung = kacarios nu calik di nagara Bituṅ Wuluṅ.

124. The personal pronouns. The following forms can be shown to be Common IN: aku, "I", kaw, "thou", ia, "he", kami, "we", kamu, "you"; in the pronoun of the third

_{* As to the use of the personal pronouns, see Essay III, §§ 118 seqq]} person plural only the nucleus ra is Common IN, the attendant articles vary, they are chiefly i or si, thus forming ira or sira.

125. The pronoun “ I ”. Philippines, Bis.: ako — Celebes, Tontb.: aku — Borneo, Day.: aku — Java, Old Jav.: aku — Sumatra, Gayo: aku — Malay Peninsula, Mal.: aku — Madagascar, Hova: ahu and zahu < i + aku (§ 44) — Northern Border, Bat.: ako — Eastern Border, Masaretese: yako < i + ako — South-Western Border, Mentaway: aku.

126. The pronoun “ thou ”. Philippines, Bis.: ikao < i + kaw — Celebes, Mak.: kaw — Borneo, Day.: ikaw — Java, Old Jav.: ko — Sumatra, Mkb.: kaw — Malay Peninsula, Mal.: kaw and ĕṅkaw — Eastern Border, Sumbawarese: kaw.

Note.—Old Jav. ko for kaw in accordance with the parallel: lod < lawd < laud.

127. The pronoun “ he ”. Philippines, Ibanag: ya — Celebes, Mak.: iya — Borneo, Sampit: iyae — Near Java, Bal.: iya — Sumatra, Angkola: ia — Malay Peninsula, Mal.: iya — Madagascar, Hova: izi < iya (§§ 18 and 24) — Eastern Border, Sumbawarese: ia — South-Western Border, Nias: ia.

128. The pronoun “ we ”. Philippines, Inv.: kami — Celebes, Tontb.: kami — Borneo, Bol.: kami — Java, Old Jav.: kami — Sumatra, Gayo: kami — Malay Peninsula, Mal.: kami — Eastern Border, Masaretese: kami.

129. The pronoun “ you ”. Philippines, Ibanag: kamu — Celebes, Tontb.: kamu — Java, Old Jav.: kamu — Sumatra, Karo, in certain districts: kamu — Malay Peninsula, Mal.: kamu.

Note.—Of all the above-mentioned personal pronouns kamu has the most restricted distribution, and accordingly we have some hesitation in pronouncing it to be Common IN.

130. The pronoun “ they ”. Philippines, Ibanag: ira — Celebes, Bareqe: sira — Borneo, Bol.: sida — Java, Old Jav.: sira — Sumatra, Toba: nasida — Eastern Border, Masaretese: sira — South-Western Border, Nias: ira. Note I.—In sida, d stands for r in accordance with the RLD-law (see § 190).

Note II.—Nias ira and ia (§ 127) only occur in certain syntactical combinations.

Note III.—Old Jav. sira is also singular.

131. The plural pronoun ra and the honorific particle ra (§§ 91 seqq.) are identical. We have a parallel in the Karo pronoun kena. This is the pronoun of the second person plural, “ you ”, without any nuance of politeness, but it can also be used in addressing a single person, and then it is polite.

132. The demonstrative pronoun. Amongst the numerous demonstrative pronouns of the several individual IN languages we may pronounce itu, “ this ”*, to be Common IN. It is a combination of the two articles i and tu.

The pronoun “ this ”. Phihppines, Bis.: ito — Celebes, Tontb.: itu — Borneo, Bol.: itu — Java, Sund.: itu — Sumatra, Bĕsĕmah: itu — Malay Peninsula, Mal.: itu — Madagascar, Hova: itu — Northern Border, Form.: iχo.

Note.—Form, iχo with χ,† as in the parallel: maχa, “ eye ”, spelt magcha, beside Common IN mata; for other examples see § 151 ad fin.

133. The interrogative pronoun. The Common IN form is apa, “ what ”, which consists of the article a and the nucleus pa.

The pronoun “ what ”. Philippines, Pamp.: apa — Celebes, Mak.: apa — Java, Old Jav.: apa — Sumatra, Karo: apa, “ anything ”, apai, “ which ” — Malay Peninsula, Mal.: apa — Eastern Border, Laoranese: apa — South-Western Border, Mentaway: apa.

Note.—Pamp. apa serves only as a word-base for forming verbal derivatives, which mean “ to go and see how (or what) a thing is ”, e.g., maṅapa. Hereto we have the parallel case that Bis. onsa, " what ", forms derivatives which are translated by " buscar, querer " (" to see k") and the like.

134. The pronoun " who " has a number of equivalents in the IN languages, but none of them can be held to be Common IN.

135. The indefinite pronoun. The Common IN form of this is anu.

The pronoun " somebody, something ". Philippines, Bis.: ano — Celebes, Tontb.: anu — Borneo, Sampit: yanu — Java, Sund.: anu — Sumatra, Gayo: anu — Malay Peninsula, Mal.: anu — Madagascar, Hova: anuna — South-Western Border, Mentaway: anu.

Note. — In Old Jav. anu has to be accompanied by the article ṅ in certain syntactical combinations, and to this anu + ṅ the Hova anuna corresponds. Thus in the Hova anuna an article has got inseparably attached at the end of the word, while in the Sampit yanu < i + anu another article has attached itself to the beginning.

Numerals.

136. The numerals are almost exclusively disyllabic formations; their analysis and the positive explanation of their component parts present great difficulties.

137. The numerals " one, ten, hundred, thousand " are Common IN in the forms sa, puluh, ratus, and ribu.

138. The numeral " one ". Philippines, Tag.: isa — Celebes, Tontb.: sa, ĕsa — Borneo, Tar.: isa — Java, Sund.: sa — Sumatra, Gayo: sa, sara — Malay Peninsula, Mal.: sa — Madagascar, Hova: isa — Northern Border, Form., Puyuma dialect: sa — Eastern Border, Sumbawarese: sa — South-Western Border, Nias: sa, sara.

139. The numeral " ten ". Philippines, Bis.: polo — Celebes, Tontb.: puluq — Borneo, Tar.: puloh — Java, Old Jav.: puluh — Sumatra, written Mkb.: puluh — Malay Peninsula, Mal.: puluh — Madagascar, Hova: fulu — Northern Border, Form., various dialects : pulo — Eastern Border., Sumbanese : kěmbuluh — South-Western Border, Mentaway : pulu.

Note. — Several IN languages, particularly in the Philipines, tolerate no h as a final. — Hova fulu follows the law : " Common IN p appears in Hova as f, save after labials or when final ". — I do not know how to explain the final q in Tontb. puluq*

140. The numeral " hundred " . Phihppines, Bis.: gatos — Celebes, Bug.: ratuq — Borneo, Tar.: ratus — Java, Sund.: ratus , Old Jav. atus — Sumatra, Gayo: ratus — Malay Penin- sula, Mai.: ratus — Madagascar, Hova: zatu — Eastern Border, Bim. : ratu.

Note. — The initial of gatos and that of zatu follow the RGH-law, for which see Conant, JAOS, XXXI, I, pp. 70 seqq. The final of ratuq follows the law: " Common IN final consonants, except nasals, become q in Bug.".

141.The numeral " thousand" Philippines, Iloko: ribu— Celebes, Tonsawang: mo-ribu — Borneo, Tar.: ribu — Java, Old Jav. : iwu — Sumatra, Gayo : ribu — Malay Penin- sula, Mai.: ribu — Madagascar, Hova: a-riuu — Eastern Border, Bim.: riwu.

Note. — In Old Jav. iwu the r ought not to have disappeared, for it was originally an r of a different shade from the one in ratus, where the r has rightly disappeared in strict accordance with the R-laws in Old Jav. (see § 190). The r in iwu ᐸ ribu ought, according to phonetic law, to have persisted, as the following table shows: Mal. Bis. Old Jav. ratus gatos atus rimbit limbit rimbit, “ to take pains ” ribu libo riwu

Parallels like rimbit show that where Mai. has an r and Bis. an l Old Jav. also exhibits an r. But the word for "hundred ""

• [See Essay IV, § 116, and also §§ 144 seqq., 185 seqq.] has influenced the word for “thousand”, and so both r′s, that of ratus and that of ribu, have been dropped. — For analogous modifications of numerals through the influence of other numerals, see §§ 33 and 183, where nomu has acquired its u from pitu.

142. There is no Common IN type for the formation of the numerals 11-19.

SECTION V: EXTENSION OF THE WORD-BASE.

Preliminary Observations.

143. In the preceding Section, §§ 81-142, we have been discussing word-bases. The term “word-base”^[18] is thoroughly appropriate and legitimate. For, in the first place, the word-bases are the shortest, and so the most fundamental, forms that have a real living existence in actual speech ; and, secondly, they serve as a basis or foundation for the further formation of derivatives.

144. Word-bases may either do duty- in a sentence just as they are, without any addition, or else they may require certain extensions to enable them to perform that task. In the Kupangese text communicated by Jonker we find the sentence: “I shall go to-morrow” = To-morrow then I go = ola kam auk laho. Here the Kupangese word-base lako is used as a predicate without any change whatsoever. In Juanmarti's Magindanao dialogues we read: “I too am well” = mikapia aku den . Here the word-base pia, “good” , has had to undergo an extension in order to fit it for serving as a predicate.

145. In works on the IN languages one often meets with the technical term " stem " (German '"Stamm"): see Misteh, “Charakteristik” , pp. 229 seqq., and Finck, “Haupttypen” , pp. 8-I seqq. But I notice that some scholars when they speak of the “stem” refer to the word-base, while others thereby denote the forms produced by extension of the word- base. And, after all, either usage can be justified, for (as has already been remarked in § 65) the elements that are used for forming the word-base from the root and those that are used in extending the latter, are in a great measure identically the same. I therefore avoid the term “stem” and speak of the word-base on the one hand and its extensions on the other.

146. As was observed in § 80, a considerable number of the formatives used in these extensions are identical with words of form. And it often happens that one and the same formative serves in the formation of both verb and substantive, and so on. Here are two problems which we cannot pursue further in the present monograph.*

The Verb.Ṫ

147. Among the verbal formatives that we find in the various individual IN languages, we can show the following to be Common IN: four active formatives: ma-, maṅ-, ba-, -um-; three passive formatives : ka-, ta-, -in-; one transitive formative: -i; and one causative formative: pa-.

148. The active formative ma-. Philippines, Magindanao: maulug, “to fall”, word-base ulug — Celebes, Tontb.: ma- sowat, “to answer” — Borneo, Day.: marabit, “to tear” — Java, Old Jav.: matukar, “to contend, to light” — Sumatra, Toba: madabu, “to fall” — Malay Peninsula, Mal.: makan, “to eat”, word-base kan, “food” — Madagascar, Hova: mahita, “to see” — Northern Border, Form.: makairi, “to work left-handed” — Eastern Border, Bim.: malampa, “to go” — South-Western Border, Nias: maliwa, “to move (one- self)”.

149. The active formative maṅ-. In most of the IN languages the final ṅ of the formative is assimilated to the initial of the word-base, so that ṅ persists only before velars and before vowels; and when the initial consonant of the word-base is a surd, that initial disappears. Accordingly in Old Jav. maṅgĕtĕm, “to pinch”, < maṅ + gĕtĕm, the final of the prefix and the initial of the word-base have remained unaffected; whereas in Tag. mamokot, “to fish”, < maṅ + pokot, an m has

• [See Essay III, §§ 35, 138.]

Ṫ [On this subject see Essay III, particularly §§ 43-117.] appeared in place of ṅ + p. — These processes also occur in the case of other prefixes, see e.g., paṅan from the word-base kan (§ 51).

The active formative man-. Philippines, Tag.: mamokot, word-base pokot — Celebes, Tontb.: mamoṅkor, " to fish ", word-base poṅkor — Borneo, Day.: maṅaput, "to darken", word-base kaput — Java, Old Jav. : manurun, " to descend ", word-base turun — Sumatra, Toba: manurat, "to write", word-base surat — Madagascar, Hova: manasa, " to wash", word-base sasa — South-Western Border, Simalurese : manasai, " to wash ", word-base sasa ; as to the -i see § 156.

150. The active formative ba-. Philippines, Bis.: baigad, " to stroke ", the word-base thereto being found in Iloko, viz. igad, " to stroke " — Celebes, Bug.: baluka, " to be loose, to be free " — Borneo, Day.: badaha, " to bleed " — Java, Sund. : baganti, " to interchange " — Sumatra, Lampong: baguna, " to be useful " — Madagascar, Hova: wawenti, " to be bulky, to be massive " — Eastern Border, Sumbawarese: basiṅin, " to bear a name, to be called " — South-Western Border, Mentaway: baliyu, " to fill ", word-base in Mak., viz. liyu, " filled ".

Note I.— Hova wa < ba follows the law: " Common IN b appears in Hova as w except after m ".

Note II.— In Lampong baguna the word-base guna is, of course, a loan-word from the Sanskrit; but the example is cited on account of the ba- and not on account of the guna.

Note III.— In Bis. and in Mentaway, formations with ba- are not numerous, so that ba- is no longer felt to be a formative, but is rather regarded as part of the word-base; that fact, however, is immaterial here, having regard to the purpose of our monograph. The disyllabic word-bases corresponding to Bis. baigad and Mentaway baliyu are no longer to be found in these two languages. We have therefore had to seek them in other languages, and analogous cases occur infra.

151. The active formative -um-. Philippines, Inv. : kuman, " to eat ", word-base kan (see § 51) — Celebes, Tontb.: kuman — Borneo, Day.: kuman — Java, Old Jav. : lumaṅlaṅ, "to roam about", word-base laṅlaṅ — -Sumatra, Toba: sumurut, "to recede", word-base surut — Malay Peninsula, Mal.: gumilaṅ, "to glitter", word-base gilaṅ — Madagascar, Hova: humana < human — Northern Border, Form. : Xumme, " to evacuate excrement ", word-base Xe, " dung " — South Western Border, Simalurese: lumaṅoy, "to swim", word-base laṅoy.

Note I. — We remarked in § 18 that the spelling of Form, words in Vlis and Happart was defective, but the striking doubling of the m in -umm-, e.g. in Xumme, is consistently carried out by them.

Note II. — Form. Xe, " dung ", stands in the same relation to Common IN tai as in the parallel case of X0, " man ", be- side the widely distributed tau.

152. The passive formative ka-. Philippines, Bont.: kalaṅo, "dried up", word-base laṅo — Celebes, Bug.: kacalla, " accursed " — Borneo, Tar. : kasukab, " opened " — Java, Old Jav.: katon, "seen", word-base ton — Sumatra, Lampong: kaděṅi, "heard" — Madagascar, Hova: hadinu, "forgotten", word-base in Pamp., viz. liṅao, "forgetful" — Eastern Border, Kamberese: kahira, " torn ".

Note. — Hova hadinu has been affected by the operation of four different phonetic laws, three of which have already been quoted; the fourth is: " Common IN final aw — for which Pamp. has ao — appears in Hova as u ".

153. The passive formative ta-. Philippines, Bis.: takiliṅ, " to incline ", word-base kiliṅ — Celebes, Tontb.: talicur, " to turn the back towards " — Borneo, Day. : tabiṅkis, "to be banished" — Java, Sund.: talaṅke, " hesitating ", word-base in Old Jav., viz. lěṅke, "slow" — Sumatra, Toba: talentes, " to stand open" — Madagascar, Hova: taburuaka, "bored through ", word-base buruaka, " hole " — Eastern Border, Kamberese: tabuṅgahu, "opened", word-base buṅgahu, "to open" — South- Western Border, Mentaway: taico, "to become visible ". 154. The passive formative -in-. Philippines, Tag. : tinawag, "called", word-base tawag — Celebes, Bulu: winwnu, " killed ", word-base wunu — Borneo, Day. : kinan, " eaten ", word-base kan — Java, Old Jav. : giněgō, " held fast ", word- base gěgō — Sumatra, Toba: tinogu, " led ", word-base togu — Madagascar, Hova: tinapaka, "broken", word-base tapaka— Northern Border, Bat.: binobun, "buried"; a cognate, though not etymologically identical, word-base hereto is Day. bumbon, " to conceal under something " — South- Western Border, Mentaway: tinibo, " exposed to the action of smoke ".

155. The signification of the active and passive formatives. We have translated the above-mentioned active and passive forms in a convenient but rough and inadequate way by infinitives and participles. In reality these formatives indicate a number of finer shades of meaning, some in one language, others, it may be, in another. And the number of these is so large that we cannot pronounce any one of them to be Common IN. The most widely spread case is the force of the formative -um- in forming the aorist.

156. The verbal formative -i. This is added to word-bases or to extensions of word-bases and makes them transitive.

The transitive formative -i. Philippines, Tag.: gaway, "to bewitch " < word-base gawa + i — Celebes, Bug. : joppai, " to tread on ", word-base joppa, " to tread " — Borneo, Tidung: taṅkubi, " to cover " — Java, Jav.: nulis, " to write ", nulisi, " to write upon, to cover (e.g. a sheet of paper) with writing " — Sumatra, Mkb.: manaṅih, "to weep", manaṅihi, "to bewail" — Malay Peninsula, Mal.: měnaṅis, "to weep", měnaṅisi, " to bewail " — Madagascar, Old Mlg. : ame, " to give presents to " (see Note) — South- Western Border, Mentaway: gagabai, " to seek ", word-base gaba or gagaba.

Note I. — Tag. gaway < gawa + i signifies " to bewitch ", the word-base gawa means " to make ". For this we have a parallel in Jav., where gawe < gawa + i means " to make ", but ma-gawe " to produce an effect by witchcraft ". Note II. — Old Jav. has a verb amah, " to hand over, to give, to give rise to an emotion ". Now the above Old Mlg. ame (which is given by Houtman) is for amah + i, with e for a + (h) + i like reni < ra + ina (§§ 24 and 93), to which is superadded the operation of the Mlg. phonetic law: " An h of other languages is not represented in Mlg.". Beside Old Mlg. ame Modern Mlg. (Hova) exhibits an ume, " to give, to give rise to an emotion ", which we have to explain as umah + i. It is true that neither in Old Jav. nor elsewhere, so far as I know, is there any such word as umah, but such a form is possible, as is shown by the parallel that in Old Jav. ajar, " to impart to ", is accompanied by a synonymous ujar. — As a consequence of its contraction ume is accentuated on the final syllable.

Note III. — Houtman's editors in the " Collection des ouvrages anciens concernant Madagascar " are of opinion that Old Mlg. ame is merely misspelt, the " orthographic vraie " being ume. But we have shown that the form ame is a possible one, and the word occurs in Houtman eight times in all, and each time spelt with an a. It would really be strange if that author had made precisely the same mistake in spelling eight times over, and in doing so had managed to hit on something sensible as well.

157. The causative formative pa-. Philippines, Nabaloi: pabunu, " to cause to be killed ", word-base bunu — Celebes, Bug. : padara, " to allow to bleed " — Borneo, Tar. : pakalap, " to make possible " — Java, Sund.: pasih, " to give ", word-base sih, "favour " — Sumatra, Angkola : pauli, "' to beautify " — Northern Border, Form.: pakiol, " to sharpen " — Eastern Border, Laoranese: padeta, "to elevate" — South- Western Border, Mentaway : pakom, " to give food to ", word-base kom," to eat ".

158. The case also occurs of one and the same word-base joined to the same formative running through many languages, so that one can declare the whole formation to be Common IN: Such a case is minum, from the word-base inum, "to drink ". "To drink", verb with formative. Philippines, Tag.: minum — Celebes, Bug. : minuṅ — Java, Old Jav. : minum — Sumatra, Toba : minum — Malay Peninsula, Mal. : minum — Madagascar, Hova: minuna — South- Western Border, Nias: minu.

159.' In many IN languages the verbal word-base unaccompanied by any formative is imperative, and this usage must be regarded as Common IN.*

Word-base as imperative. Philippines, Magindanao: sulat, " write ! " — Celebes, Mak.: lampa, " go ! " — Borneo, Day.: tiroh, " sleep ! " — Java, Old Jav.: laku, " go ! " — Sumatra, Toba: buwat, "take!" — Malay Peninsula, Mal.: paṅgil, " call ! " — Madagascar, Hova:fuha, " wake up ! " — Eastern Border, Kamberese: laku, "go!" — South- Western Border, Mentaway: ala, " take ! "

Note. — Hova has -a as a regular imperative formative (see § 30), but Richardson expressly states that fuha is also used as an imperative: the regular imperative alongside of it is fuha-z-a.

160. Specimen sentence with a word-base as an imperative : Kamberese, from the Story of the Civetcat: " Wait till we kill you ! " = Wait, (we) kill you ! = napa, mapameti kau.

161. The languages of the Philippines, North Celebes, Madagascar, and some other islands also have formatives for the formation of tenses, but the distribution of these formatives is not wide enough to entitle us to call them Common IN. The most widely spread case is that of n-, as a sign of the past tense. †

The Substantive

162. Substantives occur much more frequently without extension than verbs do. In the Banggaya sentence from the text communicated by Riedel: " We were going to the village of Seasea " = ikami ambakon do i lipu Seasea: the verb ambakon has a formative, but the substantive lipu, " village ", has not. 163. Among the substantival formatves the prefix ha- the infix -an- and the suffix -an can be shown to be Common IN. The formative ka- forms abstract nouns, -an- mostly indicates concrete things, -an denotes place.

164. The substantival fortnative ka-. Philippines, Magindanao: kaputi, "whiteness", word-base puti — Celebes, Tontb.: kawělar, " breadth " — Borneo, Day.: kagogop, " care, sorrow", word-base gogop, "worried, troubled" — Java, Sund.: kañaho, "knowledge" — Sumatra, Toba: halinu, "image produced by reflection in a mirror or in water", word-base in Old Bug., viz. lino, " to mirror oneself " — Madagascar, Hova: hatsara, "goodness".

165. The substantival formative -an-. Philippines, Bis.: tanoptop, "sound from afar", from the widely distributed root tup, which denotes various kinds of noises — Celebes, Bug.: kanuhu, "talon", beside Common IN kuku — Near Java, Mad.: sanolap, "jugglery", word-base in Jav., viz. sulap, " to juggle " — Sumatra, Toba: hanapa, " involucre ", word-base in Kawi Jav., viz. kapa, " covering " — Northern Border, Form.: kaloṅkoṅ, "talon", word-base in Iloko, viz. koṅkoṅ, "to scratch" — Eastern Border, Kamberese: tanai, "intestines", beside Common IN tai, "dung" — South-western Border, Mentaway: tanai, " dung ".

Note. — Form. l for n in kaloṅkoṅ in accordance with the parallel alak, " child ", beside Common IN anak.

166. The substantival formative -an. Phihppines, Magindanao: niugan, " coconut grove ", word-base niug, "coconut palm" — Celebes, Bug.: labuwaṅ, "anchorage" — Borneo, Day.: kayuan, "forest" - — Java, Sund.: tanjakan, "rising ground" — Sumatra, Toba: hundulan, "place to sit on" — Malay Peninsula, Mal: labuhan, "anchorage" — Madagascar, Hova: sampanana, "bifurcation", word-base sampana, " to bifurcate " — South- Western Border, Simalurese: kubaṅan, " pool where buffaloes wallow ", word-base in Mal., viz. kubaṅ, " to wallow in the mire ". 167. ln several languages we find a formative pa- used for indicating the agent, but it competes with mpa-, par-, pan-, etc., which are of course related to it, but are not identical with it; hence we cannot infer any Common IN factor here.

The Adjective.*

168. The Common IN formative for the formation of adjectives is ma-. In the Tag. riddle about the five fingers, in Starr, " Filipino Riddles " : " Five coconut trees, one is high (= higher than the others) " = limaṅ puno naṅ niog, isa i malayog: malayog is an adjective, formed by means of ma-from the word-base layog.

169. The adjectival formative ma-. Philippines, Inv.: mapia, " good ", word-base pia, " goodness "' — Celebes, Ponosakan: mapihia, "good" — Borneo, Day.: manis, "sweet" < ma + anis, "sweetness" — Java, Old Jav. : maputih, "white" — Sumatra, Toba: malimbo, "high" — Madagascar, Hova: malutu, "dirty" — Northern Border, Form.: matakot, "timid" — Eastern Border, Kamberese: maliṅu, "useful" — South-Western Border, Mentaway: mabatu, " stony ".

170. Several IN languages possess a formative for the comparative, usually -an or -ěn, but it is not distributed widely enough to enable us to call it Common IN.

The Adverb.

171. In the IN languages the adverb is mostly identical with the adjective, or it is a prepositional construction, or a substantive may be used adverbially without a preposition, and the like. Example : In Ranawaluna's Book of Laws, Article II, we find: " Theft of rice, by mowing it by night in the field " = The mowing rice (by) night there in the field= ni midzindza wari alina ani an tsaha. In this Hova sentence the substantive alina, " night ", is used without change or addition as an adverb. 172. But there are also formatives for the forming of adverbs, and among them ka-, which makes adverbs of time, is to be regarded as Common IN.

The adverbial formative ka-. Philippines, Tag. : kagabi, " yesterday ", word-base gabi, " night " — Sangir Group, Sangirese: kahěbi, " yesterday " — Celebes, Tontb. : kaawiqi, " yesterday " — Borneo, Day. : katelo, " for (i.e. during, lasting for) three days ", word-base telo, " three " — Sumatra, Mkb.: kapataṅ, "yesterday", word-base pataṅ, "evening"; kini, " now " <ka + ini, word-base ini, " this " — Madagascar, Hova: halina, " last night " <ka+alina— Eastern Border, Bim.: ka-sa-nai, " on one day ".

Note. — The Day. formula katelo, " for three days ", has its pendant in the Mlg. harua, " for two days ", word-base rua. "two".

The Numeral.

173. For the ordinals there is a Common IN formative, namely ka-. The multiplicatives and distributives also have their special formatives in the various individual IN languages, but none of them can be shown to be Common IN.

The formation of the ordinals. Philippines, Bagobo : ka-tlo, "the third" — Celebes, Tontb.: ka-tělu — Borneo, Tar.: katalu — Java, Sund.: katilu — Sumatra, Lampong: katělu — Malay Peninsula, Mal.: ka-tiga, "the third" — Eastern Border, Laoranese: kalima, " the fifth ". 174. There are, to begin with, two methods of doubling words. According to the first one the whole word is set down twice over. Thus in the Mak. children's song Daeng Camummuq there is the sentence: "Slowly, slowly swallow (the food) down your (=-nu) throat ! " = palemeq-lemeq namaqnauṅ ri kallon-nu. Here the whole word-base lemeq is reduplicated. The second case is: First the word is set down as far as the second vowel, inclusively, and then it is set down in its entirety. Thus the Day. dirge Augh Olo Balian Hapa Tiwah begins with the words: " Flee, soul of (the) dead ! " = Lila-lilaṅ liau matāy. Here lilaṅ is reduplicated according to the second method. The omission of the final consonant is not the result of any sandhi-laws of the several languages; that phenomenon is an ancient heritage. A third kind of reduplication merely repeats the first two sounds of the word-base, e.g. Bont. nonoaṅ, " toy buffalo ", beside noaṅ " real buffalo ."

The first two kinds of reduplication convey an intensification of the fundamental meaning, or, occasionally, the opposite, a weakening of it. The third kind indicates a thing, mostly a tool.

175. Reduplication to the second vowel. Philippines, Ibanag: sinnu-sinnun, " garments ", word-base sinnun, "garment" — Celebes, Tontb.: londe-londey, "all sorts of ships '— Borneo, Day.: humo-humoṅ, "somewhat stupid" — Java, Old Jav. : sulu-suluṅ, "to swarm pell-mell" — Madagascar, Hova: tiṅgi-tiṅgina, "to sit at the edge of" — Northern Border, Form.: darra-darrab, "to hue clothes thickly" —

South-Western Border, Mentaway: boli-bolit, "to twist one-self about".

176. Reduplication of the first two sounds. Philippines, Bont.: nonoaṅ, “toy buffalo” — Celebes, Bulu: tutura, “pole for pushing”, word-base tura, to push — Borneo, Day.: gagada, “vane”, beside the synonymous Mal. gada-gada — Java, Jav. : wĕwĕdi, “scarecrow”, word-base wĕdi, "timid" — Northern Border, Form.: wawarigbig, "borer", word-base warigbig, "to bore" — Eastern Border, Rottinese: sisilo, "gun", word-base silo, "to shoot" — South- Western Border, Mentaway: tutura, “pole for pushing”

Note. — The first kind of reduplication, the doubling of the entire word-base, is so very widely distributed that examples are superfluous.

---

Synthesis of Words into Sentences.

177. As already announced elsewhere, I shall publish a special monograph on this subject.^[19]

PART II

'ORIGINAL INDONESIAN

178. We saw in § 1 that the word laṅit, either unchanged or modified only in conformity with strict phonetic law, runs through a number of IN languages. How do we account for that fact ? By the assumption that there was once a uniform original IN language, which possessed the word laṅit, and from which its offshoots, when they parted away from it, took the word with them.

179. Having in § 2 styled the word laṅit “Common IN” , we now call it “Original Indonesian” , and we also apply this epithet to all the linguistic phenomena which in Part I have been pronounced to be Common IN.

180. It is self-evident that this Original Indonesian also went through a process of evolution: when we speak of the Original IN mother-tongue in this monograph we are referring to its last phase, immediately before its subdivision.

181. Indo-European research also speaks of an original mother-tongue, though with more reserve nowadays than formerly: see Meillet-Printz, p. 17, and compare therewith Porzezinski-Boehme, p. 198.

182. In the field of IN research the conditions are more favourable to the hypothesis of a common original mother-tongue. That, surely, has been proved by the whole of our dissertation on Common IN. But we will single out a few particularly striking points.

183. The several IN languages, although they extend over such an enormous area, are more closely related together than the Indo-European ones. We may illustrate that fact, for example, by the case of the numerals. We give here the numerals 1-10 of the four most outlying regions.

Northern Border	Eastern Border	South-Western Border	Western Border
Puyuma	Sumbanese	Mentaway	Hova
sa	sa	ša, šara	isa
rua	dua	rua	rua
tero	tilu	tälu	telu
spat	patu	äpat	efatra
rima	lima	lima	dimi, dima
unum	nomu	änäm	enina, enem
pitu	pitu	pitu	fitu
waro	walu	balu	walu
iwa	siwa	šiba	siwi
purru	kĕmbuluh	pulu	fulu

Note I.—The sibilant in Mentaway ša, šara, šiba and Hova siwi somewhat resembles our “ sh ”.

Note II.—Hova dima and enem occur before suffixes or in composition.

Note III.—The resemblance of dua to the corresponding Indo-European numeral is merely fortuitous; dua also occurs in languages which have no Sanskrit loan-words at all.

184. The conservative character of the IN languages is further illustrated by the way in which onomatopoeic formations run unchanged through the several individual languages. Thus flatulence is imitated by the phonetic series t + u + t, and though p + u + t or p + u + p would be equally appropriate, the nucleus tut always recurs.

Flatulence, to break wind. Phihppines, Pamp.: atut — Celebes, Tontb.: ĕntut — Borneo, Day.: ketut — Java, Sund.: hitut — Sumatra, written Mkb.: kantut — Malay Peninsula, Mal.: kĕntut — Madagascar, Hova: etutra — Northern Border, Form.: matut — South-Western Border, Mentaway: ätut.

185. Finally we will illustrate the closeness of the relationship between the IN languages by reference to an entire section of their linguistic life, viz. the adjective, comparing for that purpose the Kamberese adjective on the Eastern Eorder and the Hova adjective on the Western Border.

I. In both languages some word-bases, without the addition of any formative, may serve as adjectives: Kamb. bokul, " big "; Hova keli, " small ".

II. In both languages ma- is the chief adjectival formative: Kamb. malinu, " useful "; Hova malutu, " dirty ".

III. Both languages also use the formatives ka-, pa-, ta-, before vowels k-, p-, t-, in order to form a limited number of adjectives: Kamb. kapatan, "dark", word-base in Mal., viz. petan, " evening " ; Hova hetri < ka + etri, " growing slowly ", word-base etri, " to diminish " — Kamb. tabana, " full " ; Hova taburi, "round" — Kamb. paita, " visible "; Hova fulaka < pa + ulaka, " folded, bent ", word-base ulaka, "bend".

IV. Neither language possesses any adjectives denoting the material of which a thing consists. The substantival name of the material is simply put after the word that is to be qualified. A " stone house " is " house + stone": Kamb. uma watu; Hova tranu watu.

V. Both languages can turn the adjective into a substantive by means of the article: Kamb. na mahamu, " the good (thing) "; Hova ni marina, " the right (thing) ".

VI. After adjectives which express a state of mind the word for " mind " is added without any connecting word of form: Kamb. mahamu eti = " good + heart" = "good-hearted "; Hova afa-pu, by sandhi from afaka fu = " free + heart " = " contented ".

186. There is one IN language, and one only, that has a history, viz. Jav. The oldest phase of it is what we call Old Jav. Now it is a reasonable assumption that this Old Jav. would be particularly closely related to the Original IN. And that is really the case. By far the greater number of the phenomena which we have shown to be Common IN, and now call Original IN, are to be found in Old Jav.^[20] It is true that in the section dealing with Common IN we did not always adduce the Old Jav. as evidence, but that was merely because we wished to let the other languages of the Javanese region, which includes Bali and Madura as well, have their say too. Let us just make special mention of one item in the general agreement between Old Jav. and Common IN, or Original IN, viz. the phonetic type of the word. Old Jav. possesses the tive common vowels and in addition to them the pépét; in the interior of a word it tolerates the kapkap type of consonantal combination; at the end of a word it admits any consonant save the palatals, thus tolerating the mediæ: and these are also the chief characteristics of the Original IN type of word.^[21] — Modern Jav. has departed much further from Original IN. For example, it has turned part of its pépéts into another sound, it has given up the sulu-suluṅ type of reduplication, its passive with the formative -in- is in the act of dying out, etc.

187. Of the earlier phases of Mlg. and Bug. there are also documentary records available, though they are far less important than the Old Jav. ones. Here too we observe that the earlier phases approach more closely to the Original IN than do the modern forms of these languages. Thus in § 44 we worked back from the modern Mlg. zama to iama, and the latter form really occurs in Houtman, p. 360. And in the Old Bug. epic La-Galigo there is an expression, no longer in use nowadays, viz. amesorěṅ, "a place where one can lie down ". In § 26 we had occasion to regret that the word sor runs through only a few languages; in Old Bug. amesorěṅ < a + me + sor + ěṅ we have a fresh piece of evidence in support of it.

188. The sum total of the linguistic facts which we have shown to be Common IN and now call Original IN is quite a considerable quantity. It is true that beside these there are a good many linguistic phenomena which we could only style Common IN with hesitation, or not at all. But that is not to say that they cannot be Original IN. Bim. on the Eastern Border has a word wara, " to be, to be found somewhere ", which recurs in Nabaloi in the Philippines as guara, in Old Jav. in Java as wwara, and in Mentaway, on the South-Western Border, as bara. The original form is wara, with a w ; the other initials follow in strict accordance with phonetic law from that w. Now this word only appears in four areas of distribution, the Philippines, Java, the Eastern Border, and the South-Western Border; on our principles we cannot possibly pronounce it to be Common IN. Yet how shall we explain the fact that it occurs at these four widely separated points ? Has each of these languages created it by itself ? That would indeed be a remarkable coincidence, particularly in view of the perfect phonetic agreement. Has the word migrated ? Words with that kind of meaning are not much in the habit of migrating from one language to another; and how could it have skipped so many intervening territories ? There will be no alternative left but to pronounce wara to be an Original IN word like so many others.

Note.— Nabaloi guara < wara in conformity with the phonetic law: " Initial w appears in Nabaloi as gu ", hence also gualo, " eight " < walu. The phonetic laws which have produced Old Jav. wwara and Mentaway bara have already been mentioned.

189. The Original IN did not differ essenitially from the modern living IN languages. One important point of difference may be said to consist in the fact that it used more monosyllabic words of substance than the modern languages do. Modern Jav. has a considerable number of disyllabic words of substance which in Old Jav. were still monosyllabic: thus Old Jav. said duh, " gravy ", but Modern Jav. says duduh. Accordingly, as we go back from Modern Jav. to Old Jav. the number of monosyllabic words of substance increases; and when we go back from Old Jav. to Original IN it is to be expected that there would be a further increase.

190. The Original IN phonetic system had two distinct rs; in several of the modern languages the two rs have become fused into one; in others, again, the one r has turned into g or h, and the other r into l or d. These vicissitudes of the r sound, particularly the RGH series, have been studied in considerable detail by Dutch scholars, e.g. quite recently by Talens and Adriani for the dialects of the Talaut Archipelago; and also by Conant for the Phihppine languages.^[22]

191. In morphology some of the IN languages, e.g. Sangirese, exhibit great luxuriance, others, e.g. Bim., a slighter development, while Old Jav. occupies an intermediate position in this respect. And some such intermediate position, it may be inferred from the data set out in §§ 143-173, was also occupied by the Original IN.

1. * [See also Essay IV, § 5.]
2. [See Essay I, § 11, 1, footnotes.]
3. [See Essay I, § 11, I, footnotes.]
4. [See also Essay I. § 11, I, footnotes.]
5. [See also Essay IV, especially §§ 39 seqq.]
6. [See Essay I, § 11, I, footnotes.]
7. In Mal., rugi perhaps suggests rather a substantival sense, but in Gayo it is commonly used verbally, e.g., aku rugi, "I have suffered loss".
8. [See Essay IV, § 126, II.]
9. [See also Essay IV, §§ 200 seqq.]
10. *[But see Essay IV, §§ 307 seqq.] Ṫ [See also Essay IV, § 311, 1 ]
    ‡[See also Essay IV, § 319.]
11. [See also Essay IV, § 35, II.]
12. [See also Essay III, §§ 35, I, 37, IL]
13. [As to the use of articles before proper names, see the footnote to §91.]
14. [The original rendering here is: " der (Herr) Gross ", which illustrates better than the English translation the use of the article with a proper name (such as Be is in this context). Like German and Greek, but unlike English, Hova and some other IN languages admit the definite article before proper names of persons.]
15. The writer has no information as to the force of the ch in this word, or in Batanese in general.
    [Perhaps, like ch in Nabaloi, it is the sound rendered by c in these Essays: see Essay III, § 162, footnote.]
16. [See also Essay IV, § 144.]
17. [See also Essay I, § 90 (and footnote).]
18. [In the original, "Grundwort ": see Essay I, § 1, footnote.]
19. [The monograph here referred to appeared in 1914 under the title “Indonesisch und Indogermanisch im Satzbau”.]
20. This disposes of the hasty and premature observation in my " Prodromus ", § 8, ad fin.
21. [See also Essay IV, §§ 3, 6, 8.]
22. [See also Essay IV, §§ 40, II, 99, 129 seqq.]