Chapter 2 - Phonetic Development of Sanskrit in the Philippine languages

2.1. OUTWARD APPEARANCE OF THE LOAN-WORDS. phonetic history of Sanskrit loan-words in the Philippine(s) languages is fraught with difficulties owing mainly to the fact that there are no texts belonging to the period antedating the Spanish conquest that have come down to the present era. While texts written in the old scripts are not entirely wanting as basic materials in the present study on the phonetic development of the loan-words, they are comparatively late. 1 These texts were collected in the effort to preserve the character of the native writings, but the subjects around which these texts revolve are also late to be of significant use in the present thesis. 2 While other texts exist, e.g., the Maragtas of Panay Is- 1 The Ayer Collection from Palawan and Mindoro Islands, now housed in the Newberry Library, Chicago, Illinois. Some are deposited in the University of Michigan Library. 2 Vide Fletcher Gardner, Indie Writings in the Mindoro-Palawan Axis, and Philippine Indic Studies. Both works contain texts, transliterations and translations of specimen writings from Mindoro, Palawan, and other Philippine scripts. -131-

-132 A land, they are transliterated in the Roman script. fundamental difficulty arises in this latter instance, that is, the Spanish transliterations, which were the earliest attempts to romanize the Philippine writings, did not follow a standard system. Hence, in some or most cases, the words suffered mutilations that rendered identification almost impossible. An A word, for instance, may be taken which if it is Sanskrit is hardly recognizable - Bisaya pusunagta, "a kind of bee" Sanskrit puspaniksa, "a bee", or the widely spread word for "respect" Bikol, Bisaya, Iloko, South Mangyan & Tagalog tabi (Tagalog var. santabi) < Sanskrit ksantavya, or Bikol, Iloko & Tagalog tanikala <Sanskrit Srnkhala, "chain" and tanka, "gold". example of curious corruption is Iloko pardaya <Sanskrit parivada, "calumniating". Another is Sulu gaugari, Tagalog lagari, Iloko ragadi, Bikol lagadi, Maranao garogaji, Igorot lakachi, & Tagabili legadi, "carpenter's saw, file" < Sanskrit krakaca. With the exception of Maranao, the rest do not have cognate forms in the intervening languages. nates of the word, which, if they are Sanskrit , are also corruptions - Mal, gargaji, Javanese gargi, Sundanese gergaji, Makassar Note the cog- & Batak garagaii. The question, "Is it a Sanskrit loan-word, or is it a native word?", arises. This same question is

-133 also asked about the other words of similar nature. However, in a much broader aspect, the Sanskrit words that have been naturalized in the Philippine(s) languages retain the original structure, like Bikol ranga, "joy, comfort, bliss" Sanskrit ganiza, "diversion", Bikol himsa << Sanskrit himsa, Tagalog mukha <Sanskrit mukha, Bisaya pipila <Sanskrit pipila, Iloko rupa, "face" <Sanskrit rupa, "form", etc. There are cases where in the intervening form, the Sanskrit suffers a change. In the Philippine(s), the original morphos is retained, e.g., Malay muka >Tagalog mukha :: Sanskrit mukha, Maranao suara Sanskrit svara > Makassar sara, Malay sutera, Batak suntura Philippine(s), e.g., Tagalog gangsa :: Javanese gangsa <Sanskrit kansya, Malay gentala, "spinning wheel" <Sanskrit yantra (of. Javanese jantra), etc. In the following sub-sections, the various phenomena under which Sanskrit loan-words have adopted themselves into the phonetic system of the Philippine(s) languages will be presented. For purposes of detail, all possible instances that may come to the purview of the phenomenon under consideration will be cited. Repetition in the citation of words

-134 should not, therefore, be considered a monotony, where a word or words will be cited under two or more phenomena. 2.21. TREATMENT OF VOWELS 2.211. Sanskrit a > Phil, a. In almost all cases, the short value of Sanskrit a is retained in Philippine(s), with very few exceptions, e.g., Sulu agama ( < Sanskrit agama) shows a long value of the penultimate vowel (which, however, shifts the value of the long a-initial). Maranao and Malay agama shows "devaluation" of the initial vowel. A number of these exceptions are found in Sulu, where side by side with the shifting of the value of a from one syllable to another, Sanskrit a develops to Sulu : Sanskrit laksa, "100,000"> Sulu (sa'-laksa), "(one) ten thousands", Sanskrit rahasya > Sulu rahisa (Malay rahsia, rahasia, Old Javanese rusiya, Makassar rahasiya). Sanskrit vansa >Sulu bangsa, but Tagalog bangsa, bansa and Maranao bangensa retain the value of the Sanskrit vowel. The same is true of the intermediate forms - Malay bangsa, var. bansa, Javanese Wongsa, ³ (Sundanese, Dayak & Batak bangsa, and Makassar bansa). Note the q in Javanese, which among all the forms. shows a deviation from the common trend. 3 Journal of the Royal Asiatic Society (Straits Branch), xvi, p. 356.

-135 Sanskrit artha >Sulu arta :: Javanese & Sundanese harta, Sans parihara, "defending or preserving from (abl.)" > Sulu palihara, "to protect" :: Javanese parihara. Beside this development of Sanskrit a to Sulu a is its change to either @ or e, as in Sanskrit naraka > Sulu neraka 4 and Sanskrit pratyaya > Sulu perchava." 5 2.212. Sanskrit > Philippine(s) a. Except for a few forms, particularly in Sulu, Sanskrit @>Philippine(s) a. The exceptions are: Sanskrit >Sulu jaga (Malay jaga), Sulu bahagi, var. bhagian (of. Tagalog bahagi) <Sanskrit bhaga (bhagi-n?).6 It is worth noting that Iloko & Maranao bagi retain the a value, which is also found in Tagalog raja, var. ladya (T) and Bisaya raja, var. radya< Sanskrit raja :: Sulu rajah.7 Sanskrit sapa > Sulu sapah (Magindanao Sepa, Maranao sapa, var. kasepa), Sanskrit prakara > Sulu parkara, "case, circumstance" (Malay & Sundanese perkara, Javanese prakara, "affair"). The wide-spread term for "bow and arrow" appears to retain the value of Sanskrit a - Bikol, Bisaya, Iloko, Maranao, South Mangyan, & Tagalog pana <Sanskrit bana, "ar- ↳ Maranao naraka, Malay & Javanese naraka. 5 Malay perchaya. 6 ual. bahagi, Javanese bage, Sundanese, etc., bagi 7 Malay, Javanese Sundanese, & Batak, in various forms - raja, radia, radya, etc.

-136 row" :: Sulu panah. 8 It Similar to Magindanao sepa, and Maranao var. ka-sepa < Sanskrit sapa, is Sulu kreia, "business" < Sanskrit karya, "work". certainly has its early development in Malay keria, "profession, etc." Only a few cases of the Sanskrit > Philippine(s) a-development will be cited as it is comparatively common. It must be noted that the change does not show any preference whether the vowel is ultima, ante-penultima, or penultima. It invariably shows that the "devaluation" or loss of a mora in the Philippine(s) forms, which is apparently true in the intermediate languages, of Sanskrit a, is due to the simplicity of the phonetic system of the Malayo-Polynesian languages. 15 However, the accentuation in the Philippine(s) forms has to be discussed also (vide §2.214 infra). Sanskrit cinta > Tagalog & Pampanga sinta, by the influence of Malay chinta. Sans, liksa > Tagalog lisa, Bisaya losa, Iloko lis- 1. Tagbanua 11-us, and South Mangyan lids. South Mangyan and Tagbanua forms show a deviation from the common tendency: Sanskrit a > South Mangyan Q, Tagbanua u. It could not, however, be surmised that they may have been influenced by Bisaya losa, which has o instead of 1 in the penultima. 8 Malay panah, Javanese pana. Tagalog laho

-137- <Sanskrit rahu, which ef. with Pampanga lawo. Side by side with Tagalog bisa (< Sanskrit visa) is Bisaya lala, "serpent's poison" <Sanskrit lala, perhaps in analogy to the serpent's venom. In fig. speech, it is considered as the serpent's saliva or spittle. Sanskrit Laksa, "lac" develops in Tagalog & Bisaya as lakha, Maranao laka, Iloko & Tagalog laksa (Javanese laka). Normally, it would be expected that the Philippine(s) forms would have followed one line of development from one of the forms in the intervening language(s). But actually, the Philippine(s) forms seem to have developed from two intervening forms: Malay laksa and Javanese laka. Malay masa is the intermediate form between Sanskrit masa and Tagalog masa, "time" This form already evinces a "devaluation" of Sans a. The same phenomenon occurs in Bisaya mamsa, which though Malay mangsa takes a guttural nasal instead of the original labial nasal (anusvara), Sanskrit a in the same word, mausa, develops in Malay as a (> Bisaya a). Sanskrit devata loses the long value of the a-final in Bisaya (var. dioata), Tagalog, Maranao & South Mangyan diwata, Tagabili dewata. The loss is already noticed in Malay and Javanese dewata. Note that Tagabili deviates from the common Philippine(s) form Sanskrit > Com.Philippine(s) i - where it retains the original Sanskrit e.

-138 Isidore Dyen disputes with apparently sufficient and convincing reasons Otto Dempwolff's 10 reconstruction of Tagalog dalagall (and Batak dara, Javanese lara, Malay dara), "mai den" <°DARAH, which reconstructed form is considered by the latter to be a loan-word from Sanskrit It is unnecessary to follow up Mr. Dyen's arguments, as they do not have direct relevance to the present problem, although he goes on to point out Cecilio Lopez's 12 citation of Sanskrit daraka, "boy" as the possible etymon of the Tagalog He dismisses this on phonetic grounds - that k has never been found to have become Tagalog g (vide § 2.28). He cites nevertheless the middle Indo-Aryan (Prakrit ) daraga, "boy" as a possible etymon of the Tagalog, but dismisses the same on the assumption that "the words of other languages are from a different source and can not be compared with Tagalog" Sex is also cited as a barrier to the derivation of the Tagalog from Sanskrit He concludes that in the absence of a credible etymon, 9 Tagalog Reflexes in Malayo-Polynesian D", Language, xxiii, 3.5. 10 Vergleichende Lautlehre de Austronesischen Wortschatzes, 1.78. 11 p. Paterno (La Antigua Civilizacion Tagalog, p. 143) writes an etymology of the word, which may be termed "mythological etymology". 12¬A Comparison of Tagalog and Malay Lexicons (on Phonetico-Semantic Basis)", Bulletin No. 2, INL, 1939.

-139 ODARAH may be considered as a native Malayo-Polynesian word. Apparently, Mr. Dyen exhausts all possible points relative to the derivation of the word, but he missed one salient aspect in the Sanskrit language. The feminine form of the Pali and Prakrit , daraga, m., is daraga, f. < Sanskrit darika, f.< m., "a ., darika, f. It is likely that it is Pali or Prakrit daraga which may be the etymon of Tagalog (& S. Mang. dalaga and and Bikol daraga). It can never be dismissed that in some cases either the Prakrit or Pali forms of the Sanskrit are found to have been borrowed by Malay, Javanese, and other Malayo-Polynesian languages. 13 Furthermore, Mr. Dyen's consideration that °DARAH must be a native Malayo-Polynesian word is negatived by the probable etymon of Malay dara, "a female child", Javanese rara, woman who had just had a child" <Sanskrit forms: dara, "a girl", the long 's having been shortened in the borrowing languages, and dara, "wife", with similar shortening of the a-value. Similary, the long a-ultima in Prakrit and Pali daraga loses its long value in the Philippine(s) languages that borrow this word. It is not without reason, therefore, to consider Philippine(s) dalaga and Malay, Javanese, etc., dara to be Indian in origin. The Javanese and other forms in the intervening languages are perhaps the result of the tendency of fast". 13 cr. Sanskritsche woorden in het Bisaya, p. 285, Re derivation of puasa, "to

-140 these languages to dissyllabism. 14 It is evident that Malay diala had influenced the development of Tagalog & Pampanga dala, "fishing net, cast net", from Sanskrit jala. Old Javanese dayah, "young woman of high rank" shows a shortening of the long a's of Sanskrit jaya, "wife", from which Old Javanese borrowed the word. It is certain that this Old Javanese form may have been the intermediate form between Sanskrit and Malay dayang & Tagalog dayang, Sulu daiang, all mean d The CJav., "lay-in-waiting, court lady, mistress, etc.". Malay, Tagalog & Sulu (with the exception of a-ultima) forms show a loss of the long a-value of the Sanskrit form. With no citable intervening forms, Sanskrit tyaga loses the value of the a-penultima as it becomes Tagalog tiyaga, var. tyaga. If Tagalog gangsa, var. gansa, ganza, Iloko gansa (Igorot kang-kangsa) were Sanskrit kansya, with Javanese gangsa as the intervening form, then its change may be explained in terms of the development of Sanskrit a to Philippine(s) a (with other phonetic developments.) 2.213. Other developments of Sanskrit a, a. Note has been taken above of the unsual development of Sanskrit a, a in Sulu perchaya <Sanskrit pratyaya, Sulu neraka < Sanskrit naraka, Maranao ka-sepa, Magindanao Sepa <Sanskrit sapa, and Sulu kreia < 14 cf. the development of Sanskrit stri to Javanese estri.

-141 Sanskrit karya. The development of a to e in Sulu perchaya and kreja is an influence of Malay perchaya and keria (the latter being a pepet vowel). the latter Malay word as karia. However, Abbe Favre renders The in Sulu neraka and e in Maranao ka-sepa and Magindanao sepa do not have cognate developments in either Javanese or Malay The same is true of e in Sulu tenga, "lead" <Sanskrit tankana, Other e formations are found in Sulu lebi, "to exceed" which is influenced by Malay lebeh, "more than" (Sanskrit labha. In Sulu preksa and Maranao periksa are seen both Malay E and pereksa, var. preksa. Three isolated developments of a and a are seen in Iloko sari, "to break through" (sara, "horn") << Sanskrit Sara< Vari, "to rend, to destroy" (>sara, "arrow"). In Bisaya & Tagalog tumbaga, which seems to be influenced? by Sundanese, Makassar, and Batak tombaga (Javanese & Malay tembaga), shows an u development from Sanskrit tamraka, "copper". The initial a in Sanskrit kacchapi becomes u in Tagalog kudyapi and Bikol kudbut shows another form in Malay which has e - kechapi. The other development of a in Sanskrit Kacchapi is found in Bisaya kodyapi, Iloko kodiapi and Maranao kotiapi. It is difficult to ascertain the origin of the development of Sanskrit , a to o in Philippine(s) Whether it is an influence of accentuation or dissyllabism, it is not known. As it is yape,

-142 irregular in some words, its ascertainment can not yet be made, except perhaps in ohly one case: Iloko iloko, "the name of the region, the language, and the people" <Sanskrit loka, "district". The final in Iloko seems to be Sanskrit ah-ending in declension, which when it follows a word with an a-initial is reduced to a while the a-initial is replaced by an avagraha. It is however difficult to say whether or not Iloko (with a prothetic i, vide § 2.24) is a Sanskrit loan-word.15 An unusual development of Sanskrit a to o is seen in Bikol gapos, and Igorot kapos, "cotton" <Sanskrit karpasa. No intermediate form is citable, except Minihasa, a minor Malayo-Polynesian These two developspeech, which has kapeh and Achinese gapes. ments: a> and @ are exceptions to the phonetic formations of Sanskrit >Malayo-Polynesian a, particularly in karpasa. Sanskrit ghatika, "time, duration of time", has its initial a developing to o in Sulu kotika (yeto), "when at the time" and Maranao kotika, "season, moment". Magindanao kutika, "time", shows an u development. Apparently, a does not undergo change in the intervening languages Malay & Dayak -katika, Batak hatika. The a's in Sanskrit danda, "punishment, 15 cf. Sanskrit gna >?Tagalog, Bikol, & Bisaya ginoo; Tagalog & Bisaya ginang (vide fn. 76, p. 32). Sanskrit a-ultima >?Philippine(s) o and ng.

-143 fine", show three developments in the Eastern languages. First, Toba Batak dandan retains the a-value of the original. Second, Malay denda and Javanese denda show both a pepet vowel and a normal e in the penultima. And third, Iloko dondon manifests an o for both the a's. Graphically, Sanskrit a, in this case, would be represented as: Sanskrit a-initial > Toba Batak a, Malay 9, Javanese e, and Iloko o; and final a>o. The penultimate a in Sanskrit ganda >o in Bisaya gonda, "disaster". No intervening form can be found in Malay or Javanese is 2.214. Accentuation and long a in poly- and dissyllabic loan-words, This a phonetic phenomenon in Philippine(s) languages where the acute accent (°), which is a stress on the vowel, is always pronounced with (1) an apparent rapidity when it is found in a final-vowel and apparently possessing the value of a final aspirate, (2) a raising of the voice occurs when it is found in either initial, antepenultima or penultima positions. E.G., Tagalog gupit; "cut with scissors" > gupitan, "barber shop", Tagalog simba, "hear mass" > simbahan, "church", Tagalog hiraman, "lending institution" Sanskrit accent, udatta.

-144 The relation between the accent in Philippine(s) languages, particularly in the penultimate vowel a in dissyllabic words, and the long a in Sanskrit loan-words, more specifically in the initial a, is apparent on a closer study of both the borrowed and the original forms. This also evident in the intervening forms of these borrowed words. It is likely that Sanskrit a is represented in Philippine(s) languages, in this particular case, by the accent. The value or quality of the a>a can not be found to be merely disguised, for the fullness of this vowel in Philippine(s) speeches appeared to be doubtful. It sounds as if it were a mora shorter, but closer scrutiny of the vowel-value gave the lie to this doubt. Illustrations. Sanskrit ayus, "life, etc." >Iloko ayus, lit. "life breath". Cognate forms are not found in Malay and Javanese Sanskrit asa>Tagalog, Bikol & South Mangyan asa. Note that the a-ultima loses its long value in the Philippine(s) form, like Malay asa. Sanskrit naga, meaning "serpent" and "tree" > Bikol and Maranao naga, "serpent", and Tagalog & Bisaya naga, "tree". Sanskrit pana, "to drink">Bisaya pana, "id.". Tagalog & Bisaya maya, "poet., "the bird as piece de comparaison concerning passing fancies or illusions" <Sanskrit maya. But, of. Magindanao masa, "time" <Sanskrit masa, "time" <Sanskrit masa, "month, moon"

-145 Tagalog lasa, "delight" <Sanskrit lasa, "pleasures". Cf. the penultimate a in Sanskrit rasa, "taste", which develops to a in Tagalog lasa, "id." and appears to be an exception to the common tendency. Cf. furthermore, Sanskrit vatsa, "child or boy, offspring" > Bisaya, Dibabaon Mandaya, Sulu & Tagalog bata and Maranao wata, oata, "id.". Sanskrit tara >Bikol & Tagalog tala, Iloko (kamonta)tala. A preliminary note on the Iloko may be made here. The Iloko lexicons do not explain the first part of the compound kamon-. The first ta shows a reduplication of the initial syllable of the base. With kamon-, it appears to be a prefix that expresses the idea of likeness or similarity. The Iloko grammars do not show this "prefix" to be in use. For a tentative explanation of this formation, however, the following may be advanced. The lexicons record kamontatala to mean "Lucifer, the morning star, venus", but it appears to mean more precisely "the likeness of a star, like It may be something that twinkles like a star, the morning dew that is struck by the rays of the a star". e.g., early sun. Cr. Old Javanese tara, "star". In contrast to the common development of Sanskrit vayu in Tagalog bagyu, Bikol & Iloko bagyo and Dibabaon Mandaya bagiu, Bisaya accents bagyu in the penultimate vowel, thus retain-

-146 ing the value of Sanskrit a. In the polysyllabic Sanskrit anyaya, the penultimate a is accented in Tagalog anvaya, which shows an unchanged value of the phoneme. Cf., however, Bisaya anava, "careless in deeds", which accents the a-penultima, but in the original Sans, anaya, "bad conduct", the a-penultima is but one mora in quality. 2.215. The a-, a-ultima. Beside the unaccented and quasi-visarga quality of Sanskrit a-, a-ultime, when they are naturalized in Philippine(s) speeches, two other developments occur. These occur in either polysyllabic or dissyllabic loan-words. The first of these developments is the unaccented final a, a with glottal catch or hamzah (\) or (2)7, A few examples may be cited: Tagalog suka, "vinegar" <Sanskrit cukra. Cr. Sulu sukak', which to all appearance is an influence of the Arabic pure hamzah. Tagalog bala, "threat" <Sanskrit bala, "strength", Sanskrit sala, "rampart" > Tagalog sala, "dam", Sanskrit kubia > Tagalog kuba. Polysyllabic examples are also citable under this phenomenon. Sanskrit vartta > Bisaya & Tagalog balita, South Mangyan barita, South Mangyan & Tagalog palibhasa, "underestimate" < Sanskrit paribhasa, "correction" The intermediate forms in Malay

-147 perbasa, peribahasa, "maxim" do not appear to have influenced this development. Sanskrit sahasa, "violence" > Tagalog gahasa, "rape", and South Mangyan sagasa, "enforce". gahasa, It may be remarked that the final a, a with the hamzah influences the quality of the penultimate vowel, whether a, u, or i, which if it does, shows a long phonetic value, as it has been seen in both the dissyllabic and polysyllabic examples. The second development is the accented final a, a with the hamzah (^). Sanskrit siddha, "accomplished" > Tagalog sidha, "esteemed"; Sanskrit vaha, var. vaha, vaha >Bikol, South Mangyan & Tagalog baha; Sanskrit visama > Tagalog Sama, and Sanskrit sutra > Tagalog sutla are few examples for occurrences in dissyllabic words. There are a number of polysyllabic words possessing accented final a. E.g., Sanskrit paduka > Tagalog paruka, Sanskrit Carita > South Mangyan & Tagalog salita, Sanskrit alocita >Tagalog alusitha, and Sanskrit dhrta >Tagalog, Bok., & So. Hang, dalit. The last example shows an accented antepenultima a. Except for dalita, the possible accent in the penultima or ante-penultima in both the dissyllabic and plysyllabic words is assimilated in the ultima, e.g. vide above, and Sanskrit putana >Bisaya putana.

-148- 2.216. 1, I. Sanskrit i whether ultima, penultima or ante-penultima is treated like Sanskrit a in the Philippine(s) languages. The same is true in the intervening speeches. Sanskrit saksin is unlikely the ancestor of Bikol, Igorot, Iloko, Tagalog & South Mangyan saksi, "witness, testimony". It is, however, very likely that the word is Sanskrit saksi (> Malay saksi). much Lit. Javanese has carita, which shows a complete naturalization from Sanskrit carita. This readily develops in Tagalog as salita, "word, tale, etc." and South Mangyan salita, "sobbing".. The Sanskrit i suffers no change in Philippine(s), as it may also be seen in Sanskrit sirisa, "mimosa sirissa">Tagalog siriCf. Iloko sarisa, which though retaining the I-penultima, shows a change in the i-ante-penultima to a. The 1> a development is also seen in Sanskrit ahi > Tagalog ahas and Sulu haas, and Sanskrit linga >Iloko langa. sa. More examples of Sanskrit 1> Philippine(s) i. Sans dhuni > Bikol, South Mangyan & Tagalog huni, Bisaya honi, Iloko & Dibabaon Mandaya uni, and Maranao oni; Sanskrit hiusa >Bikol himsa; Sanskrit visaya > Philippine(s) bisaya; Sanskrit upari >Philippine(s) Aparri; Sanskrit linga >?Philippine(s) lingayen. The epenthetic -yen in the last word will presently be remarked upon. Sanskrit 1, in both ultima and penultima, suffers change. It loses its value, and it is unaccented. Sanskrit

-149 kasturi > Tagalog kastuli, Bisaya katsuli, Iloko kastoli (Malay kesturi); Sanskrit kartari > Bisaya katli (cf. Iloko kartib). While Bikol renders the I to e in kudyape <Sanskrit kacchapi, Bisaya Iloko, Maranao & Tagalog have i as in Malay kechapi. Sanskrit nirgundi > Bisaya & Tagalog lagundi (Javanese legundi, Malay legundi). If Tagalog si is Sanskrit sri, then the change is already observed in Malay seri. Sanskrit Vasanti, "name of various plants" >?Bisaya balanti, "plant", which, however, shows rather improbable phonetic changes. The I in the penultima already undergoes change in the intervening forms, e.g., Malay & Old Javanese sigra <Sanskrit SIghra >Tagalog sigla; Dayak pariksa, Javanese priksa <Sanskrit pariksa >Tagalog paligsa, Maranao periksa and Sulu, in formative namariksa praksa. Sanskrit I has an accented development which follows the same principle as the accented formation of Sanskrit a, a. It occurs in the penultima more specifically in dissyllabic, but rarely in polysyllabic words. Sanskrit Iva >Tagalog, Bisaya, Bikol & South Mangyan diva. Sanskrit hina >Tagalog, Bikol & South Mangyan hina. Cf., however, Sulu hina which has an unaccented I (and a long value of the final vowel). The development of Bikol, South Mangyan & Tagalog is influenced by Javanese and Malay hina. The i-penultima in polysyllabic Sanskrit pipila shows an accented value in both Bisaya forms hapila and pipila.

-150 But the 1-penultima, which is an isolated case, becomes accented in Iloko Sarita <Sanskrit carita. Tagalog salita. Cf. South Mangyan and 2.217. Sanskrit u>Philippine(s) u. u. u, o, a. To cite a few of the unchanged Sanskrit u in Philippine(s) would, it is believed, suffice to illustrate this phenomenon. E.G., Sanskrit manusya >Tag, manusia; Sanskrit mukha > Tagalog mukha (& Central Cagayan Negrito mukaat); Sanskrit putri > Bisaya putli; Sanskrit Suci >Tagalog & Pampanga susi; and Sanskrit dhuni> Bikol, South Mangyan & Tagalog huni, Iloko and Dibabaon Mandaya uni. This unchanged value may be accounted for the retention of the phoneme's value in Javanese & Malay manusya; Malay muka; Javanese putri; Old Javanese mukha, etc. Sanskrit u, whether or not it is influenced by its development in Malay or Javanese, becomes u in Philippine(s) in both the penultima and ante-penultima positions. E.G., Sanskrit puja > Bikol, Bisaya & Iloko pugay, Sanskrit bhuti > Tagalog buti, Sanskrit bhumi > Sulu buni (Malay & Javanese bumi), Sanskrit Sula, "any sharp instrument for impaling" >Bisaya & Tagalog sula, Sanskrit karcura, "kind or turmeric" >?Bisaya gansuli, "plant", Sanskrit cudamani (cuda, "crest, crown" - mani, "jewel") > Tagalog Sula, "precious stone", & Bisaya sula, "crest of a sugarcane plant", and Sanskrit putana >Bisaya (sang) putana(n), "gloom, doom". Javanese putih, if it is Sanskrit , may be Tagalog & Pampanga puti "white" <

-151 Sanskrit bhuti, "white ash", Of., however, Malay putih, "white, clear colour, blank, purity", which is derived by Abbe Favre from Sanskrit puti, "purity". While either Javanese or Malay does not show a development of Sanskrit , to the rounded back vowel o, the Philippine(s) languages show this change side by side with the u-development. May this be explained in terms of the confounding of u and o in articulation, which is evident in the written language? 16 Beside Tagalog mula and Iloko mula, there is Maranao mola <Sanskrit mula, Maranao oni, Bisaya honi beside Bikol, etc. huni, Iloko uni <Sanskrit dhuni, Maranao potri :: Bisaya put li, Javanese putri <Sanskrit putri, South Mangyan moka :: Tagalog mukha, Malay muka, Old Javanese mukha <Sanskrit mukha, Maranao manosia :: Tagalog manusia, Javanese manusya <Sanskrit manusya, Maranao garoda :: Javanese garuda <Sanskrit garuda, but Malay geroda, Mal, gola :: Bisaya gula- (man), "sweet jelly" <Sanskrit gula, Maranao bodiman :: Tagalog & Bisaya budhi <Sanskrit buddhiman, buddhi, Maranao morka <Sanskrit murkha, Maranao lasona<Sanskrit lasuna, var. lasona, Tagalog laho, Pampanga lawo :: Mal, lahu <Sanskrit rahu, Maranao goro :: Sulu gu ru, Javanese & Malay guru <Sanskrit guru. The Tamil u in vilangu, although it does not change in Malay belenggu, changes in 16 In the early Philippine system of writing, the u and o are used interchangeably.

-152 Philippine(s) except Tagabili belanggu: Iloko, Maranao, Tagalog & South Mangyan, bilanggo. In the foregoing list of illustrations, it is noticeable that Maranao seems to invariably change the u to o in contrast to the common u-development in the other Philippine(s) languages. It is interesting to note also that the Maranao lexicons examined do not record u as one of the letters of the alphabet. Sporadically, is found beside Q. Two instances are found where Sanskrit u assumes the long value in Sulu: guru <Sanskrit guru and guna <Sanskrit guna. An isolated instance of a-development of Sans is citable; Philippine(s) aparriSanskrit upari. . u in Philippine(s) The accented Philippine(s) u<Sanskrit u, u in the penultima position appears, like the other two vowels, a, i, to have a value similar to Sanskrit u, although it has already been remarked that it is quite indistinct owing to the raising of the voice in stress. A few of this occurrence may be cited: beside Sulu guru, is Tagalog guru, Iloko mula <Sanskrit mula, Sanskrit dhupa > Tagalog dupa (dupa-an, Malay dupa), Sanskrit guha >Tagalog guho, Sanskrit samuha >Bisaya samuha :: ?Tagalog salamuha. 2.218. The semi-vowels and other vocalic changes. The semi-vowels undergo a variety of change as words possessing these are borrowed in the Philippine(s) The y is invar-

-153 iably retained. It, however, assumes its consonantal value: Sanskrit anyaya > Tagalog anyaya, Sanskrit tyaga >Tagalog tyaga var. tiyaga, Sanskrit udaya > Iloko daya, Sans, vayu >Philippine(s) bagyu, bagyo, etc. In a number of instances, I reverts to its vocalie equivalent, by the influence of the intervening forms. Apparently, Malay rahasia, rahsia influenced the formation of Sulu rahisa (<Sanskrit rahasya). Old Javanese rusiya and Makassar rahasiya show a svarabhakti vowel, like Tagbanua bagiyo<Sanskrit vayu. Sanskrit chava > Sulu chahia (cf. Igorot chaia, which shows i instead of y. Makassar chaya). While Tagalog dayang retains y by the influence of Malay dayang, Sulu daiang changes y in Sanskrit iaya to i. Sans manusya >Tagalog manusia and Maranao manosia (Malay manusia), Tagalog santabi and other Philippine(s) forms, tabi, syncopate the final vowel (except Bisaya tabia), and the semi-vowel reverts to its vocalic value. It is palatalized in a few cases: palatal media Sanskrit karya > Sulu kreja by the influence of Malay kerja (karya, Favre); the aspirate palatal tenuis by the influence of the dental tenuis - Sanskrit pratyaya > Malay perchaya> Sulu perchaya, but Tagalog gantala shows a guttural where Malay jentera and Javanese jantra show a palatal media

-154 to yantra. It is unusual for Tagalog have developed the palatal to guttural where it normally naturalizes the phoneme to a dental d semi-vowel y as in radya, ladya <Sanskrit raia. Cf. Bisaya radya, Malay radya or radia. Mod. Tagalog raia, but i no longer has the palatal value, but the aspirate value due to the influence of Spanish 1, which sounded as h (i takes the place of Spanish h which is silent). is The phoneme is either lost or retained in Sanskrit vaniiva >Tagalog banig, Iloko banyaga, Bisaya & Dibabaon Mandaya baligva. But it entirely disappears in Iloko palangka <Sanskrit paryanka or palyanka, probably by the influence of Malay pelankin var. pelanki (cf. Oriya palanka, Ardha Magadhi pallanka), Bisaya sutaSanskrit cyuta, Bisaya samala <Sanskrit Syamala. The Philippine(s) languages do not have y in their phonetic systems. Two distinctive developments occur to this phoneme in Philippine(s) The first and most common is its development into the labial plosive media, and it is generally seen in the initial position. But, it rarely occurs in the intervocalic. A few examples may be cited: Sanskrit vansa > Sulu bangsa, Tagalog bansa, Maranao bangensa, Sanskrit vans > Tagalog & South Mangyan bangsi, Sanskrit vrata > Bisaya balata, etc., Sanskrit vrana > Bisaya balana. 17 17 cf. this development with Sanskrit y> Javanese b in bayu < vayu, bangsa

-155 The intervocalic position of the plosive is seen in Tagalog santabi and other Philippine(s) forms <Sanskrit ksantavya. The development of Sanskrit v in vismaya to Ilk, d in dismaya is out of the question, for the Iloko is Spanish desmayo. It The other development is its appearance as W. apparently retains its vocalic value when intervocalic, either in its original position or when anaptyxis or svarabhakti occurs. In the first instance, Sanskrit avah > Tagalog & South Mangyan awa, Sanskrit jIva > Malay jiwa > Tagalog, Bisaya, Bikol & South Mangyan diwa, Sanskrit devata >Philippine(s) diwata, dewata, and perhaps Sanskrit saindhava > Tagalog sanyawa, Sulu sendawa and Maranao sandawa (Malay & Javanese sendawa). In the second instance, there is Sanskrit sva > Philippine(s) asawa, except Sulu which has asava. It is, however, strange that Sanskrit Y > Sulu v, while beside this is Sans v>Sulu w (vide supra). Cf. Bisaya gami "master", which H. Kern 18 derives from Sanskrit svamin. There is an isolated case where, although the semivowel reverts to its vocalic equivalent, the phonetic value is exactly like the original Sanskrit svara > Sulu suala and Maranao suara. Here, an anaptyctic vowel might have been expected as in Sanskrit sva >Philippine(s) asawa or a 18 Sanskritsche woorden in het Bisaya, p. 282.

-156 W-consonant as in Malay svara, Javanese & Batak sowara. Isolated changes of Sanskrit y are found. Sanskrit upavasa > Sulu & Bisaya puasa, where it also reverts to its vocalic equivalent after the syncopation of a, which is unusual, and where it is expected to be retained. The other development is its assimilation by the cerebral? Sanskrit Svarga Sulu shagra, "heaven" and shegra, "the sun". The latter development in Sulu could have been expected to develop along the same line of change as in Javanese or Malay suwarga (Makassar & Sundanese surga). Malay puasa, Makassar puwasa. Puasa, however, has forms in Sanskrit I presents an equally varied development which except in one case does not have intermediate forms in Malay19 thetic Sanskrit Sara South Mangyan sira and Iloko sari (ef. Iloko sara, "horn"), Sanskrit arp "to creep"> Bikol & Iloko sirip (Tagalog silip). It develops into a liquid, e.g., Sanskrit dhrta > Tagalog dalita, Bikol dalita and South Mangyan dalita (darita). It develops into a front vowel u Sanskrit pariprsta > Tagalog & South Mangyan alipusta and Sanskrit rsya > Philippine(s) usa, usa, ugsa, uisa and ugtag (Bikol osa). Philippine(s) usa, etc, has a form in Malay rusa. 19 cr. the development of the phoneme in Javanese - Sanskrit krta >Javanese kerta or karta, nrpati > prepati, nrapati, bhrtya >breta (Favre, Gram. Jay., p. xii-xiv).

-157 Incidentally, it appears that the semi-vowel I assumes both its guna and/or viddhi forms in the Philippine(s) developments. Note for instance, Iloko sari and sara and Tagalog dalita (<-ar-), the accent seems to show a vrddhi value of the semi-vowel. The phoneme that appears after the liquid in dalita and perhaps in silip or sirip is a svarabhakti vowel. The 1- and a-ultima in Iloko sari and sara are more or less epenthetic developments. The i in ir- and il-developments may have been due to the gloss of the (>ri, re, re) in phonetic articulation. The u-form is unexplainable. Whether Iloko & Tagalog tanikala is Sanskrit , or more broadly an Indian loan-word, it is difficult to say. If it is an Indian loan-word, two elements are apparent tani and kala, which are from two Indian words Sanskrit tanka, "gold", as suggested in the second meaning of Iloko "golden chain" and srnkhala, "chain". It is obvious that the final syllable of the first word and the initial of the second have been dropped, thus tan-khala. The i in the Philippine(s) form would then be anaptyctic, and kh is de-aspirated. The has a A Philippine(s) var., talikala, the first element of which is analogous to or an influence of Philippine(s) talf, "rope, twine, string". Other vocalic changes are incidental. The au in Sulu gaugari and e in Tagabili legadi are two developments of

-158 the a-initial in Sanskrit krakaca. Even the intermediate forms can not provide proofs to its being a loan-word from Sanskrit Malay gargaji, Sundanese gergail, etc. Sanskrit a, i, Sanskrit pusks > Bisaya u, a, at, are by no means widespread. paniksa > Bisaya pusunagta, "bee". H. Kern 20 explains that su (so) in the second syllable in Bisaya must originate from (pu)swa, in which p is weakened to u, and that i in the penultimate syllable must have been replaced by a dull e (pepet). Then, when an accent was added to the same syllable, it became a. e is the Malay equivalent of Philippine(s) a. 21 Hence, the word may have already undergone a considerable change in Malay Taking this as a starting point, H. Kern's explanation of the development of the second syllable may be examined. The first part of the compound has a cognate in Malay, which shows a very slight change from Sanskrit - puspa, "flower". It is doubtless that the second syllable sa developed from spa, after p has been dropped (for no discoverable reasons) and a> by the influence of the preceding vowel. But ks > gt is an explicable phenomenon (vide § 2.27. >). 20 Sanskritsche woorden in het Bisaya, p. 285. 21 The pepet vowel is also found in Philippine(s) languages. Vide C. E. Conant, "The Pepet Law in the Philippine Languages.

-159- 22 2.22. Treatment of Diphthongs. e. If Iloko sipat, "clapping of hands", is not native Philippine(s), it may be Sanskrit Gapeta. Sanskrit > Iloko a. Where the diphthong in Sanskrit devata becomes common i in Philippine(s) divata, it is retained in Tagabili dewata. Bisaya, Iloko & Tagalog bisti, "dress, garment", may be Spanish vestir, "to dress", vesti, "dress", but it also shows resemblance with Sanskrit vesti, "a thing tied around, meaning particularly the cloth or dhoti". Both have which in Philippine(s) becomes 1. It is difficult to decide from which language the Philippine(s) has been borrowed. Spanish may claim the credit for its source, for the language has also enriched the Philippine(s) languages. But, there is the probability that its ancestry is Indian. The e in Sanskrit lekha, "ray, line", lekha, "god, deity", apparently shows that it is the gun or i in Vlikh. It further shows a reversion to the simple vowel as it is borrowed by Philippine(s) - Tagalog 11 ha, Pampanga & Iloko 11-a; Tagalog & South Mangyan likha, "icon". Sanskrit megha shows at least three developments in Tagalog & Pampanga 1579 first, Tagalog magha, with var. bigha, megha; second, Tagalog var. megha is Malay mega; and third, Pampanga biga appears to show a development from the other Tagalog var. bigha, with the loss of h. 22 cf. Latin vesta > vest.

-160- 9. Invariably, Sanskrit o becomes Philippine(s) u, with a very few exceptions. These exceptions may be found in Sanskrit kostha > Bisaya Iloko, Maranao & Sulu kota :: Tagalog kuta (cf. Sanskrit kostha Tagbanua kuta). It is apparent that Malay kota has influenced the more widespread or common form of the word. Cf., however, Javanese kuta and Batak huta. Another of these exceptions is Sulu drahka var. dahulaka <Sanskrit drohaka, beside Tagalog duluhaka. The intermediate forms (Malay derhaka, Javanese duraka) do not provide any clue to the Philippine(s) formations. Sanskrit koti > Tagalog kati, "ten millions", but Malay keti, "one hundred millions". The more common development is seen in Sanskrit alocita >Tagalog alusitha, "investigation"; Sans, codaya, "to importunate" > Bisaya sudaya, "admonish"; Sanskrit parusa >Tagalog parusa, Tagbanua parusaan :: Javanese parusa, var. prosa, prusa "using violence", 23 etc. Like e, Sanskrit o in dosa, moksa, lobha, and paribhoga, bhu-pari). is the guna of u (in //dus, /lubh, 1/2. muc. and With the exception of Maranao dosa and Sulu var. dosa, dosah, 23 Dyen (op. cit.) writes that parusa does not have Sanskrit parusa as its ancestor, for the form is a compound with the prefix padusa, with lingualization of the d when intervocalic. The assumption is, perhaps, negatived by the presence of Sanskrit parusa in Old Javanese, parusa, which is likely to be the ancestor of the Tagalog, and Tagbanua parusa-an.

-161 the o>u-formation is regular, which shows that the guna vowel in the original word reverts to the simple vowel in the Philippine(s) languages Bikol, Iloko, Sulu & Tagalog dusa, dusah (cf. Malay & Javanese (osa). Tagalog muksa (Old Javanese muksa, "to disappear, perish"), Tagalog lubha, Bisaya malubha, lubha, Bisaya malubha, "delirious", and Tagalog alibugha. ai. With only two Sanskrit words - saindhava and maithi. 11, the development of the diphthong shows no less than four forms in the borrowing languages of the Malayo-Polynesian family, two of which are found in the Philippine(s) group. Sanskrit saindhava, "a kind of rock salt" > Sulu sendawa, Maranao sandawa, Tagalog sanyawa. These developments in Philippine(s), except Tagalog, are perhaps influenced by Malay sendawa or/and Javanese sendawa. Cf. Sundanese chindawa, which shows another development. The ai in maithili becomes a in Bisaya mantili, and shows that it is a cognate of aixa-development in Sanskrit saindhava. It is also notable that ai being the vrddhi of 1 shows an apparent guna value in Sulu sendawa and Javanese sendawa, while Malay has a pepet vowel. P au. The diphthong in Sanskrit kaumudi develos to a in Sulu kamudi. elopment. Malay kamudi may have influenced this dev-

-162- 2.23. Anaptyxis or Svarabhakti. What appears to be the basic principle around which this phonetic phenomenon revolves is that the svarabhakti vowel is influenced by the vowel in the vicinity of the consonant cluster, either in the preceding or following position. However, where the vowels in the word are the same, the question of which would the anaptyctic vowel be does not arise. If it does, it is found in a very few cases. Tagalog bahala may be Where the loan-word is dissyllabic, the tendency is that the initial vowel influences the svarabhakti vowel. Sanskrit bhaga (bhagin) >Tagalog, Magindanao & South Mangyan bahagi, Sulu bahagi, by the influence of the neighbouring Malay bahagi (beside Iloko & Maranao bagi, Javanese bage). Sanskrit bhara, beside Bisaya bala. It appears, however, that Tagalog bahala has its earlier form in Malay bahara, "weight". Sanskrit vrata > Bisaya, Pampanga & Tagalog balata. Maranao baratapa< Sanskrit Otapa, Sanskrit vrana >Bisaya balana, Sanskrit sranta >Tagalog salanta, Iloko saranta. 24 Sanskrit drohaka > Tagalog duluhaka, 24 In his attempted rectification or reconstruction of Tavera's derivation of Tagalog si, "an honorific particle placed before names of persons" <Sanskrit sri, F. R. Blake ("Sanskrit Loan-Words in Tagalog", JHUG, xxii, 163, p. 65) writes that the Sanskrit would naturally develop into sali from the consonantal group sr being represented by sal as in salanta <Sanskrit sranta. This assumption would violate the rule where the neighbouring vowel influences the avarabhakti vowel.

-163 Sulu dahulaka, beside Malay derhaka, Javanese duraka; Sanskrit grahana > Maranao garahana. sva The monosyllabic Sanskrit Sva assumes a polysyllabic morphos as it is assimilated into the vocabulary of the Philippine(s) languages, with final a influencing the syarabhakti vowel: Igorot, Iloko, South Mangyan & Tagalog asawa, Sulu asava. There appears to be no intermediate form of the word. There are, apparently, exceptions to the supplementary rule, for it is the vowel of the ultima which influences the anaptyctic vowel: Tagalog gantala <Sanskrit yantra <Malay gentala, jentera, Javanese jantra, and Bisaya, Pampanga & Tagalog mantala Sanskrit mantra Malay mantera, Javanese mantra. Note, however, the e in Malay jentera and mantera. In the first anaptyctic development, Sanskrit sampratyaya shows a similar tendency as those words with conjuncts for initial phonemes. But, in the second process, it is either from the preceding or following vowel that the svarabhakti vowel is developed. However, this rule appears It would more or less be like Malay seri <Sanskrit sri if si is Sans, in origin, where e in Malay is certainly a gloss of i. The analogy between Blake's reconstruction and the Tagalog salanta <Sanskrit sranta would not stand scrutiny for the svarabhakti a in salanta is influenced by the neighbouring a, thus sranta >s-a-lanta. Furthermore, his reconstruction would only be applicable to salanta and not to the word that is being reconstructed. Cr. Sulu siripada Sanskrit sri pada, in which the anaptyctic vowel is influenced by the neighbouring I.

-164 to be inapplicable to some words borrowed from Sanskrit What would have been expected from the development of Sanskrit sampratyaya to Tagalog was sampalatiyaya instead of sampalataya, in which the semi-vowel would play the role of influencing the anaptystic vowel. Example of this expected development are Sanskrit tyaga > Tagalog var. tiyage (beside common traga), Sanskrit vyaya >Tagalog & South Mangyan biyaya, Sanskrit abhyasa >South Mangyan & Tagalog bihasa (Sulu biaksa). The last two examples may have been influenced by Malay biyaya, biasa and biyasa. Beside the development of Sanskrit yayu in Tagalog bagyu and Bisaya bagyo, there is in Tagbanua, bagiyo, which shows that the svarabhakti vowel i is influenced by the semi-vowel y. The vowel 1 appears to be the anaptystic vowel where no trace of its existence in the Indo-Aryan source is found, e.g., Sanskrit vartta > Bisaya, South Mangyan & Tagalog balita. Bikol bareta shows e, which is usually confounded with i. The development of the semi-vowel (>-ar, -al, -ir, -11) has already been discussed in §2.218. The vowels that come after this development may be considered epenthetic (final) in cases where the word is dissyllabic with no consonantal ending, e.g., South Mangyan & Iloko sari and sara also < Sanskrit Ver. Tagalog sira <Sanskrit Ver; But in the case of

-165 dissyllabic words with consonantal endings, e.g., Bikol & Iloko sirip, Tagalog silip <Sanskrit srip, and in polysyllabic developments of dissyllabic loan-words, e.g., Tagalog dalita <Sanskrit dhrta, the 1 is svarabhakti. The 1 may well have been confounded with e, e or i, because of the gloss of the I in phonetic articulation. 2.24. Prothesis. As in anaptyxis, the influence of the neighbouring vowel is felt in prothesis. There are two instances in this phonetic phenomenon where the influence of the vowel in the vicinity is seen, e.g., Igorot, Iloko, South Mangyan & Tagalog asawa, Sulu asava and South Mangyan & Tagalog aksaya <Sanskrit sva and ksaya, respectively. In aksaya, a svarabhakti vowel would have been expected. The i in Iloko iloko appears to be prothetic comparable to the 1 in Tamil, which is prefixed to loan-words beginning with a lingual or liquid, e.g., Tamil irama <Sanskrit rama, "the name of the epic hero in the Ram.", Tamil ilangasogam < ?Malay langkasuka, "name of one of the territories of Sri Vijaya conquered by the Cola emperor, Rajendra, in the 11 th century".25 Prefixed to names of places, i assumes a genetive or ablative function, e.g., Illano (Bulletin de l'Ecole Francaise d'Extreme Orient, xviii, 6, p. 11) writes on Ilangasogam, "L'i'initial fait ne pas difficulte, le Tamoul ajoutant souvent un i au debout de mots etrangers commencant par une linguale ou une liquide..."

-166- 26 "people of Lanao" or "..from Lanao", " 26 Ibanag (Old Javanese tan madoh i sor, "down there". 27 Tagalog has ibabaw, "on the surface", etc., Buginese, iliwen, "on the other side", Malagasy (Madagascar) i masu, "before the eyes," etc. 28 It may not be out of place if the etymologies of Iloko iloko will be further discussed here. John Crawfurd 29 writes that Ilocano is derived from Tagalog iloc, "river". Similarly, A. A. Chamberlain 30 Chamberlain 30 thinks that iloko is Tagalog iloc. Ilocano is the Spanish form of the word with the suffix 26 vide Wilkes, op. cit., in The Philippine Islands, xliii, p. 181, fn. 70. 27 Brandstetter, op. cit., p. 43. 28 Ibid. 29 Dictionary, p. 155. 30"Philippine Studies I - Place Names", American Anti quarian, xxii, 6, p. 396.

-167 ano or no, masc. meaning "he who is from or of Iloko origin, that which is from or made in the Iloko region: cloth, rice, etc." Should The form iloc has never been found to be of this orthography in the Tagalog lexicons and in the Iloko dictionaries. Tagalog ilog, "river" is that which is found as well as Iloko ilog, "creek (of salt water), obs. river". The question of the final Q, and intervocalic k arises. Iloko iloko be Tagalog ilog, after eliminating the corrupt form iloc, is there any internal phonetic phenomenon whereby g>k on becoming intervocalic on the suffixation of Q, which appears to be more or less a Spanish influence? The o as Spanish influence may be ruled out of the question because it shows other features which will be discussed presently. But then, Iloko, itself, has ilog, which could not have suffered change when o was suffixed. g and kin both Iloko and Tagalog show independence even in intervocalic position Iloko bakara, "a kind of small fish found at the mouths of rivers," abaga, "shoulders", boko, "knuckles, node", Tagalog kakana, "story", kagat, "bite" (Iloko kagat), Iloko & Tagalog kugon, "a species of the grass family", kuko, "finger nail", abaka, "a fiber plant", etc. This may justify its inclusion in §1.39.11, Place Names with Sanskrit origins, The development of o will be -

-168 discussed in § 2.26. The In 2.25. Epenthesis: Spontaneous Nasalization. 31 phonetic history of Sanskrit loan-words in the Philippine(s) languages shows divergent developments, and therefore, difficulties in identification would be present. Some of these words show morphological changes, like the appearances of an epenthetic nasal within and at the end of the word. some cases, where compounds of both Sanskrit and native Malayo-Polynesian words are found, a nasal appears to link the two words. Both Sanskrit words are also found to be linked up by this phoneme. Sometimes the nasal disappears, and in two isolated cases a syllabic final epenthesis occurs. This development appears to be peculiar only in the Philippine(s) languages. In the subsequent pages, the various epenthetic developments will be discussed extensively. Medial nasal epenthesis. Beside Bikol & Iloko saka, 32 "the legs, branch of a tree" ( < Sanskrit Sakha) is the widely spread term for "branch of a tree" Bikol, Iloko, Dibabaon Mandaya 31 cf. Gonda, op. cit., p. 233, et seg. 32 In Natya Sastra (Art of Dance), in one of the kinds of Abhinaya, the phase of dancing or mimetic expression as it spreads or develops from limb to limb of the body is technically called sakha, the name being based on the imagery of the limbs of the body or the branches of the tree (Vide Bharata Natya Sastra, viii, 7).

-169- 33 South Mangyan, Pamp & Tagalog sanga, Igorot panga, Tagabili hanga. While J. Gonda 34 attests H. Kern's 35 observation that n in Bisaya mantili served to disguise the Sanskrit maithili, the present writer prefers to take it with caution. However, H. N. van der Tuuk 36 thinks that Javanese mantili is Sanskrit maithili, which perhaps attests that Bisaya mantili has an Indo-Aryan origin. If it is, then Bisaya mantili may be derived through Javanese mantili. The development of the nasal in Bisaya is already known in Javanese turmeric". Bisaya gansuli, "a plant" Sanskrit karcura, "a kind of Then may be considered, however, to have replaced the lingual. This phonetic development does not show any similar development in the other Philippine(s) languages or in the Malayo-Polynesian languages in the Indonesian group. Where two words either both native Philippine(s) or Sanskrit native Malayo-Polynesian the epenthetic nasal appears. Tagalog laka-n ( <Javanese raka, vide §1.322), when compounded with other native Malayo-Polynesian words, e.g., Malay bini, "wife", Tagalog bakor, "fence", the nasal appears. Thus, lakanbini, or lakambini, 33 cr. Malay anggota, "limbs "member" 34 Ibid., p. 234. 35 Sanskritsche woorden in het Bisaya, p. 284. 36 Journal of the Royal Asiatic Society, xiii, 1881, n.s. arms, legs" <Sanskrit anga,

-170- "the consort of the Tagalog deity - lakanpati ( <Sanskrit pati, "lord, protector")", lakanbakor, "the deity of the Tagalogs who is responsible for the protection of the fields". Note particularly the Tagalog compound word lakanpati. If the second part, pati, is a foreign word, it may be Sanskrit as suggested. Its being compounded with the native word lakan ( < Javanese raka) showed the epenthetic nasal. It may also be true of Tagalog amansinaya ( Sanskrit sri nava (Sri Nayaka), "lord, guide, leader"), "the deity propitiated by the Tagalogs for a good catch in fishing". Cf. Old Javanese tarangana <Sanskrit tara gana. Final nasal epenthesis. The n-ultima in Iloko dondon (<Sanskrit danda) may have its early usage in Toba Batak dandan, whose influence did not affect Malay denda and Javanese denda. Beside Tagalog kuba, Iloko kubbo and South Mangyan kuba is Bisaya kubong, which shows the final nasal ( <Sanskrit kubja). Malay dayang, shows a final nasal, which apparently is the intermediate form between Sanskrit jaya and Philippine(s) Tagalog dayang and Sulu daiang. Two doubtful place names show the epenthetic n-ultima Palawan Sanskrit pallava (vide fn. 245, p.117), and Kamalaniugan Sanskrit kamala - yuga. Loss of the nasal. Side by side with Iloko kasumba, which retains the labial-nasal, are Tagalog kasubha and Bisaya

-171 kasubha, which show a loss of the nasal ( < Sanskrit kusumbha). Malay kesumba and Javanese kesumba retain the nasal. If Iloko & South Mangyan supot were a Sanskrit loan-word, then it may show a loss of the nasal from samputa. Other modifications of the nasal. Invariably, the Sans, cerebral-nasal becomes dental in Philippine(s), which already undergoes change in the intermediate languages. Sanskrit Sans. Sanskrit bana > Philippine(s) pana, panah :: Malay panah, Javanese pana, ganda > Bisaya gonda. It may be noted that d>d in Bisaya guna >Sulu guna :: Malay guna, Sanskrit grahana >Maranao garahana :: Malay gerhana. The Sanskrit palatal-nasal n followed by a guttural tenuis, e.g., tankana, becomes a guttural nasal in Philippine(s), e.g., Tagalog tingga, Sulu tenga. A corollary development also occurs to the guttural tenuis: it becomes a media. This development already occurs in Malay tingkal, but in which the tenuis is retained. Perhaps, the development of the anusvara to either the dental or guttural nasal in Philippine(s) is explained by its change in the intervening languages. Sanskrit vamsa > Sulu bangsa, Tagalog bansa, bangsa, Maranao bangensa. In these examples, the anusvara shows two distinct forms, the 1 and ng (n). It may be surmised that Tagalog form with n may

-172 have been influenced by the Malay var. bansa. The other forms, Sulu bangsa and Maranao bangensa are developed by the influence of the common Malay bangsa (:: Batak, Dayak & Sundanese bangsa, Javanese wongsa). The anusvara ng is also found in Sanskrit vans > South Mangyan & Tagalog bangsi, via Malay bangsi, and Sanskrit hamsa > Philippine(s) gangsa :: Malay hangsa. Cf., however, Bengali bankasala <Sanskrit vaniivasala > Iloko bangsal :: Malay bangsal. Other consonantal epenthesis. This phonetic phenomenon in the Philippine(s) does not have external influence, except perhaps in one instance. As in the nasal epenthesis, it may either be final or medial. Iloko padak, "footfall", paddak, "sole of the foot, footprint" <Sanskrit pada. The epenthetic k may have been suffixed in analogy with tep-tep, "to knock (a sheaf of paper, etc.) against something (to put in order)", pek-pek, "the sound of wood being beaten". Sanskrit mani >Sulu manik-manik, Tagalog manik³7 and South Mangyan manik. This word is the exception referred to above. It may have its early form in Malay manik. The phonetic phenomenon in this particular Malay form may have occurred in Sanskrit mani and is not probably due to 37 cr. B. B. Schneider, guage", Philippine Journal of Science, vii, 3, p. 174. nik-manik, 'bead "Notes on the Mangyan LanTag. & prob. Sambali, "ma-

-173 Tamil manikkam, "ruby". Malay has also a mani, meaning "semen", derived from Arabic mani, "semen", and probably to distinguish mani, "gem or bead" from mani "semen", k may have been added to the former word. -k may then be later than the 14 th century, the probable date of the incursions of the Arab language in the peninsula. However, of, Sanskrit manika, "jewel", which may be the ancestor of both the Malay and Philippine(s) forms. The q' (-k) in Central Cagayan Negrito dayaq', "east" (<Iloko daya) is a development which is inexplicable in the phonetic history of the language. ence of the widely spread consonantal endings which tend to be stops, glottal or otherwise in quality, in Iloko, which is a contiguous language. CONeg. mukaat, "face" (:: Tagalog mukha, South Mangyan mukaSanskrit 38 It is comparable to or an influThe epenthetic y in Philippine(s) panday, or its vocalic equivalent in Sulu pandei may be an influence of Malay panThe y in Iloko turay <Sanskrit tura, in Bisaya & Tagalog pugay Sanskrit puja (and perhaps in Tagalog talisay <Sanskrit tali sa) do not show analogous developments in the other Philippine(s) dei. 38 cr. Munda (Santal, Khasi, Sora) and Jap. muk, "to face", Chinese, mok, "eye" Bisaya muklat, "to open the eyes" (vide Kaufmann, Bisaya-English Dictionary, p. 593).

-174 languages or in the intervening speeches. 39 The other epenthetic final consonant, -1, has no analogies in the other languages. Iloko & Tagalog palakol, if it developed from Sanskrit parasu shows the -1 epenthesis (vide § 2.27, for more discussion). If Philippine(s) bagyu, bagyo, baguio, or bagiu be derived from Sanskrit Yayu, its development is expected to be influenced by Malay bayu. The appearance of a where it is not expected can not be explained. Similarly, the appearance of -1- in Pangasinan ulse, "deer", is unexpected. However, it may be a metathetic development of Malay rusa, with r>1 change. Syllabic epenthesis. It is with caution that this section is appended. However, by its being included here, a new light may be thrown upon this seemingly doubtful development. The place name, Lingayen, "the place of the linga" according to H. O. Beyer, 40 if it has Sens. linga for an ancestor, would meet no difficulty in explaining its syllabic ending. Apparently, it is a locative ending. 39 cr. Sanskrit a-final > Tamil ai, e.g., Sanskrit puja > Tamil pujai, Sita >Sitai, panda > pandai, etc. 40 Philippine Journal of Science, 1 xxvii, Item No. 2.

-175 Bisaya gulaman <Sanskrit gula and Maranao bangensa <Sanskrit vansa are other examples of this phenomenon showing -man and en-final and medial epenthetic syllables, respectively. The sa in Iloko rupsa Sanskrit rup may also be a native accretion, like si in Tagalog tulasi <Sanskrit tula. ever, si is already known in the intervening languages (vide fn. 44, p. 15). HowThe as- 2.26. Treatment of the Aspirates and H. pirates kh-gh, th-dh, and ph-bh generally lose the h when words possessing these sounds are naturalized in the Philippine(s) This already occurs in the intervening languages.41 kh-gh. Sanskrit sakha > Bikol & Iloko saka, "foot, branch of a tree", has no form in Malay (but of. Old Javanese saka, "post, column" <Sanskrit sakha, "branch, door post"). Sanskrit sankha does not only lose h in South Mangyan songgo, but also changes the tenuis k to its media equivalent, & (cf. Malay sangka). Sanskrit sighra > Tagalog sigla. This word is found already de-aspirated in Malay segera and Javanese sigra. ch-ih. These aspirates show very irregular developments. A var. of Bisaya kodyapi, that is kotsapi, appears ⚫ to retain the quality of the palatal in Sanskrit kacchapi. 41 Vide Abbe Favre, Grammaire Javanaise, Introduction, Sanskrit aspiree > Javanese non-aspiree. p. xii-xiv.

-176 The sound is preserved in Malay kechapi. The other Philippine(s) forms have, like the common Bisaya kodyapi, -dy-< -ech- development: Tagalog kudyapi, Bikol kudyape, kodyapa (or kadyapa, "a green plant "< Sanskrit kacchapa, "cedrela toona, L"). -di- is seen in Iloko kodiapi, while -ti- is shown in Maranao kotiapi. Whether or not Sanskrit Vmurch, "to rouse, to excite", is the origin of Sulu moeka, "rage of a raja", it shows very interesting developments. The palatal assumes a guttural value, while the lingual is replaced by the diphthong e, which is rather unexpected. The word, however, may have been borrowed through Malay or Javanese murka, "wrath of god or a ruler" from Sanskrit murkha, but still showing the r >?e. th-dh. When Sanskrit artha is borrowed by Javanese the h undergoes metathesis, hence harta, but it is lost in Sulu arta. Sans, athava, "either or" > Sulu atau, "either- or". The development of this word is already seen in Malay atau, "id.", beside its var. atawa (and Javanese atawa, var. utawa). It may be remarked that beside the loss of h, the final a is syncopated, which influences the vocalization of the semi-vowel y. h in Sanskrit maithill is lost in Bisaya mantili.

-177 Similarly, dh undergoes a general de-aspiration as words possessing the sound are borrowed in Philippine(s) languages. Sanskrit dhara loses its aspirate value in Philippine(s) Tagalog dal, Bisaya dala, "bear a burden", Bikol dara, "to carry", and South Mangyan dala, dara; Sanskrit dhupa >Malay dupa > Tagalog dupa; Sanskrit buddhiman >Malay budiman >Maranao bodiman; Sanskrit sandhi > Iloko sandi; Sanskrit Suddhi, "purity" (?Malay sudi, "satisfied"; Old Javanese sudi, coll. "being willing", obs. "benevolent") > Iloko sudi, "brilliance" (cf. ?Tagalog surhi, "to certify", Bisaya sudhi, "blame", which retain the value of the aspirate). Sanskrit saindhava > Malay sendawa, Javanese Sendawa > Sulu sendawa. Of. Sundanese chindawa, which shows a metathetic h. The h of dh in Sans, maharddhika is lost, and d apparently is assimilated by the liquid in Tagalog maharlika (vide §2.28, re RLD Law in Philippine(s) languages). Beside Cur. Malay merdeka, "independent", is merdeheka, which shows a svarabhakti pepet vowel (Mod. Javanese merdeka, id.). ph-bh. Perhaps the only Sanskrit word possessing the labi al-aspirate sound is Sanskrit phala which becomes Tagalog pala, "gain by" (Malay pala, "fruit"). Tagalog shows a deaspirated morphos, and perhaps is influenced by the Malay form. With the exception of Sulu bhasa, the initial labial aspirate media is de-aspirated to its simple consonantal

-178 equivalent, b, in the other Philippine(s) form, basa <Sanskrit bhaSans. bhanga > Bikol, Bisaya & Tagalog bangga, banggi, 8 a. which is also found in Old Javanese banga, "refractory, rebellious", Mod. Javanese, "to resist (the police)". Beside Bikol & Bisaya bahandi, which shows an anaptyctic a, are Pampanga & Tagalog bandi Javanese banda, "riches, capital, funds", Malay benda, "things, article"). Sanskrit bhuti > Tagalog buti and Sanskrit bhumi > Sulu bumi :: Malay & Javanese bumi. Beside Bis, & Tagalog kasubha, which retains the aspirate is Iloko kasumba <Sanskrit kusumbha. The intermediate forms, Malay lebeh, Old Javanese lowih and Javanese luwih, show metathetic h, where Bisaya & Tagalog laba, Sulu lebi lose it (<9 ans. labha). h appears. The Where the appearance of his very much unlikely, it is found in Sanskrit alocita becoming Tagalog alusitha, and Sanskrit chaya developing into Sulu chahia. former has no form in either Javanese or Malay, while the latter has Malay chahaya. The appearance of the second h is perhaps an influence of the aspirate. Cf. Javanese chahya. In Sulu, where it occurs almost naturally, the appearance of h as final may be explained either as a development in the intervening languages or as a phenomenon that has occurred in the Indian locus (e.g., the visarga,

-179 which usually appears as regular ending "of a word in pausa for a or " and this occurs in declensions 42). r Sanskrit guda >?Sulu gotah, var. gatah :: Malay getah. Beside Com. Philippine(s) dusa, which shows no similar appearance of h in ultima, is Sulu dusah, var. dosah. Sans • bana > Malay panah >Sulu panah; Sanskrit raja > Sulu rajah (cf. Malay raja); and Sanskrit sapa > Sulu sapah :: Malay serpa, serapah. Metathesis of the h. In three cases, h undergoes a change of position. The bh in Sanskrit bhattara and paribhoga becomes de-aspirated and his later found with t and 6, respectively :: Bisaya, South Mangyan & Tagalog bathala 43 and Tagalog alibugha. Pampanga batala loses h. Cr. Malay batara and Javanese bhatara. h in Sanskrit argha becomes initial in Bisaya, So. Mang,, Sulu & Tagalog halaga, halga :: Maranao arga. The h-initial is already observed in Malay harga. Cf. also Sanskrit drohaka > Sulu dahulaka :: Tagalog duluhaka, where h is seen to undergo metathesis in Sulu. This metathesis of h, particularly in the labialaspirate media, bh, is not expected for the consonant normally admits a svarabhakti vowel (vide § 2.23). 42 A. A. Macdonnel, A Sanskrit Grammar for Students p. 4, fn. 1. 43 For early etymological studies of this word, vide The Philippine Islands, x 1, p. 69, fn. 23, and Paterno, op. cit., p. 36-39, P. 36-33 de 44-47, and 50-52.

-180 Like Tagalog dupa (< Sanskrit dhupa), it is expected that Sanskrit dhuni would merely lose h, instead it is d - Bikol, South Mangyan & Tagalog humi, Bisaya honi. Iloko uni loses the aspirate, for like Pampanga, it does not admit h in its phonetic system. Similarly, Sanskrit lekha, "line, ray", would have developed in analogy with the same word meaning "god, deity" Tagalog likha, "icon". But, it is entirely changed Tagalog liha, Iloko li-a, etc., "each division of an orange fruit". - The development of h (Sanskrit hamsa) to g in Philippine(s) (Bisaya & Tagalog gansa, Bikol, Iloko & Maranao ganso) may be explained in terms of its development in Malay gangsa (vide §1.22). When the aspirate is retained or when it appears with another consonant, either tenuis or media, in the process of metathesis, the aspirate quality is no longer retained. It is bifurcated in the process of syllabication, unlike in the original word where the aspirate constitutes a single syllable with the following vowel. Sanskrit ka-tha > South Mangyan, Sulu & Tagalog kat-ha, Sanskrit duhkha > Bikol & Tagalog dukha, etc. This follows the principle in Philippine(s) languages that where two consonants are contiguous in a word, syllabication is always effected at the point of contigu-

-181 The visarga in two citable cases disappears in the borrowed word Sanskrit punah, "new, again" >Bisaya puna, "to repair, make new". The visarga, however, appears to have been merely replaced by the accent in the Bisaya word. manah >Maranao maana :: South Mangyan maan, which perhaps shows a metathesis of the a-final. Sanskrit The o in Iloko (& Igorot) diso and iloko is likely to have been developed from a visarga in Sanskrit disah and lokah, which in sandhi becomes o by the influence of the following a, which in turn is replaced by an avagraha. But, the possibility of a direct borrowing in this case is out of the question, for there is in Malay loka, var. loga, ar., "place, region", and Javanese loka, "id.", which show a loss of the visarga, if the word is Sanskrit lokah. & Malay desa. Metathesis of other phonemes. syllable pra, particularly I and a, prakara, undergo metathesis in Sulu Cf. Javanese The phonemes in the in Sanskrit pratyaya and perchaya and parkara. This development has its origins in the intervening language: Malay perchaya and perkara (cf., however, Javanese pra- 44 Vide L. Bloomfield, Tagalog Texts with Grammatical Analysis, pt. 2, 25, p. 138-139. The Author uses muk-ha, "face" <Sanskrit mu-kha) as one of his illustrations.

-182 kara). Sulu rahisa<Sanskrit rahasya (Malay rahsia, rahasia, Old Javanese rusiya, Makassar rahasiya). The Sulu form may have its origin in Malay rahsia, which in turn has developed from the older fora rahasia, whose medial a has been syncopated. s and 1 in Malay are only interchanged in position to form Sulu rahisa. An anaptyctic i appears between s and y in Old Javanese and Makassar, while the semi-vowel assumes its vocalic equivalent. Beside the normal Sanskrit kasturi > Tagalog & Iloko kastu- 11, kastoli, and Sulu kasturi, a positional change of the -st- conjunct is seen in Bisaya katsuli. 2.27. Treatment of Sibilants. The simplicity of the Philippine(s) phonetic systems makes the process of borrowing less complex. For in its simplicity, the words adopted into the IN languages, within which the Philippine(s) speeches are classified in the vast Malayo-Polynesian family, the "sounds must accommodate themselves to the phonetic capacities of the languages that accept them".45 Thus, a clear-cut manifestation of R. Brandstetter's postulate, particularly as it affects Philippine(s) languages, has been seen in the previous pages. 45 Brandstetter, op. cit., Essay V, xii, 281-289, p. 327, et sea.

-183 The phonemes under study, whether dental, palatal or cerebral, develop invariably to dental (or ? quasi-palatal as they are accommodated into the Philippine(s) systems. There are a number of exceptions which will presently be discussed including the quasi-palatal development of these sibilants, Sanskrit s > Philippine(s) s. Sanskrit asa >Tagalog, Bikol & South Mangyan asa; Sanskrit sala > Bisaya, Iloko & Tagalog sala; Sanskrit Sakuna > Tagalog sakuna; Sanskrit sighra >Tagalog sigla; Sanskrit pranasa >Bisaya punasa, etc. An exception to this common tendency is citable. S becomes a guttural tenuis k Sanskrit parasu > Iloko & Tagalog palakol, Bisaya pakol, if the Philippine(s) is not native Malayo-Polynesian word. But the Philippine(s) forms may yet turn out to be Sanskrit kola, "a weapon". Cf. Old Javanese parkola, "hatchet, axe", which if Sanskrit kola is its ancestor shows that the initial syllable par- may merely be a local accretion. It is likely then that the Philippine(s) forms may have their early origins in Old Javanese parkola. It shows very interesting phonetic changes as it was adopted in the Philippine(s) languages. The change would be Sanskrit kola > Old Javanese parkola with a syllabic prothesis > Iloko & Tagalog palakol, with syncopation or metathesis of of the final a or furthermore a svarabhakti a appears. -

-184 A liquification of the lingual occurs. 46 Sanskrit s > Philippine(s) s. Sanskrit visa >Bisaya, Maranao, Sulu & already shows change in the intervening lanTag. bisa. guages Javanese & Malay bisa. A similar phenomenon occurs in the journey of Sanskrit dosa and paribhasa to the Islands - Bikol, Iloko, Maranao & Tagalog dusa, Sulu dusah; South Mangyan & palibhasa (Javanese & Malay dosa, Malay perbasa, peribahasa). Sanskrit visaya > Philippine(s) bisaya, Sanskrit Sirisa >Iloko & Tagalog sarisa, sirisa, Sanskrit dusta > Bikol & Tagalog dusta, Sanskrit pariprsta > South Mangyan & Tagalog alipusta, etc. No exceptions are citable under this phonetic change. Sanskrit S Sanskrit s > Philippine(s), in various developments. svarga shows a ?palatal or ?dental development in Sulu shagra, "heaven", and shegra, "the sun", beside Maranao sorga (Malay & Javanese suwarga, Makassar & Sundanese surga). Sanskrit sahasa shows a g-development in Tagalog gahasa, beside South Mangyan sagasa, if (again) the Philippine(s) is not native Malayo-Polynesian (cf. Tagalog var., Bikol & South Mangyan var. dahas). Should Bisaya balanti be Sanskrit vasanti as H. Kern 47 has suggested, then the liquid development of the dental sibilant would seem strange, 46 Cf., however, C. Douglas Chretien, "The Dialect of the Sierra de Mariveles Negritos", U. of Calif. Publ. in Linguistics, iv, 2, p. 75. The root pakol seems to be evident in Iloko & Tagalog palakol, with the infix -al- (-el-), which is common in the IN languages. 47 Sanskritsche woorden in het Bisaya, p. 287.

-185 particularly as it occurs only in the Philippines. Sanskrit s in saksi becomes t in Pangasinan tasi and Tagbanua t-um-aksi. s is lost in Tagalog Fahuyo Sanskrit rahasya. Sanskrit ks > Philippine(s) s in initial, and ks (gs) in intervocalic positions. Sanskrit kantavya >Tagalog santabi. It would, however, be expected that either phoneme would be dropped out, or a svarabhakti vowel would have appeared between them. The other example is Sanskrit ksipa > Bikol, Maranao, Sulu ic Tagalog sipa, sipa. It may be assumed that this latter example has its earlier development in Malay, if it is not native Malayo-Polynesian word sepak, "football". tion. There are a number of examples illustrating the ks (gs) development of Sanskrit ks in the intervocalic posiSans. moksa > Tagalog muksa (Old Javanese muksa), Sanskrit laksa > Sulu laksa, Bisaya, Iloko, South Mangyan & Tagalog laksa (Javanese & Malay laksa), Sanskrit saksi > Bisaya, Iloko, South Mangyan & Tagalog (Igorot), saksi, Tagb t-um-aksi (of. Pangasinan tasi); Sanskrit pariksa > Tagalog paligsa :: Sulu preksa and Maranao periksa (Malay pereksa, Javanese priksa). This development of Sanskrit ks > Philippine(s) ks or gs is demonstrably attested by its earlier change in Malay (and Javanese). Beside this common development, there are others which are worth noticing here. s disappears but h takes its place: Sanskrit laksa, "lac">

-186 Tagalog & Bisaya lakha, beside Maranao laka. k disappears: Sanskrit liksa > Tagalog lisa, Bisaya losa, South Mangyan lios, and Iloko lis-a. Apparently, the pausa or caesura in Iloko may have taken the place of k. Ks becomes gt in Bis, pusunagta< $ Sanskrit puspanika. 2.28. Other Consonantal changes. Sanskrit k, g>Philippine(s) either k or g, and vice versa. Discussion on the development of Sanskrit k to Philippine(s) g, from §2,212, particularly in reference to Sanskrit darika, daraka> Philippine(s) dalaga, may now Isidore Dyen had dismissed that Sanskrit k of 48 be resumed. any source has become Tagalog g. Dismissing even the example of daraka > dalaga, Dyen's belief is negatived by the widespread development of Sanskrit k to Philippine(s) g, side by side with its retention as Philippine(s) k. Sanskrit Karpasa > Bisaya, Dibabaon Mandaya To cite a few examples: & Maranao gapas, Bikol gapos :: Sulu kapas and Iloko kapas. In the Malay-Indonesian group, there is Achinese gapes. Sanskrit tamraka > Bisaya & Tagalog tumbaga :: Javanese & Malay tembaga. Beside Tagalog sampaka are Bikol, & Iloko sampaga <Sanskrit campaka. Sanskrit loka>Malay loka, var. logia, which shows a g-development in the var., while 48 vide R. Pischel, Grammatik der Prakrit-Sprachen, 202, p. 145-146. Sanskrit k when medial is found already to become Prakrit & (?which latter may represent Old Javanese g). Vide also Sanskritsche woorden in het Tagala, p. 259. Vide, furthermore, Subhadra Jha, Comparative Grammar of the Prakrit Languages, p. 149-150.

-187 Iloko iloko retains the original sound. Sanskrit vikala Tagalog bigal, "weak", Sanskrit rikta > Tagalog lista, beside likta (according to Tavera), Sanskrit sakala >Bisaya sagala, "fragment, piece". Sanskrit kansya > Tagalog gangsa, Iloko gansa, etc. Sanskrit g is lost in Iloko langa, while it is retained in Philippine(s) lingaven <Sanskrit linga. Sans, c. i>Philippine(s) s. dy (di), d (r), s. Sanskrit C shows a sibilant (dental?) development in Philippine(s), except in few or isolated cases. Sanskrit cakra > Bisaya & Tagalog sakla, but ef. Malay chakera, chakra (Favre), Sanskrit candana > Bisaya, South Mangyan & Tagalog sandana :: Malay chendana, Sanskrit carita > South Mangyan & Tagalog salita, Iloko sarita :: Malay cherita, cheritera, Sanskrit cinta > Tagalog & Pampanga sinta :: Malay chinta, Sanskrit cukra > Philippine(s) suka (Sulu sukak', Maranao soka, Pangasinan tuka) :: Malay chuka, Sanskrit kunci > Tagalog & Tagabili kunst, Maranao gonsi :: Malay kunchi, Sanskrit vicara > Tagalog bisala, but Sulu bichara :: Malay bichara, Javanese wichara, Sanskrit marica > Bisaya malisa, Maranao marisa :: Malay mericha. One salient point that may be noted in the foregoing illustrations of the phenomenon is that it appears that Sanskrit c> Malay & Javanese ch did not influence an equivalent development in the Philippine(s), except Sulu. This may be due to the proximity of Sulu to

-188 Malay (or Javanese). Sanskrit @ shows a ty form in Bisaya katya > Sanskrit kaca, beside Tagalog kasa (T). Cf. this, however, with Sulu kacha, and Malay kacha, both having developed also from Sanskrit kaca. and chh have a peculiar development (vide §2.26). Apparently, C, chh in Sanskrit are taken as one in Philippine(s), whereby both phonemes are fully assimilated in the forms dy, di, or ti Tagalog kudvapi, Bisaya kodyapi, Bikol kudyape, Iloko kodiapi, and Maranao kotiapi. Bisaya has a var. which more or less approaches the quality of Sanskrit c, chh kotsapi. kacchapi. It is probably influenced by Malay kechapi < Sanskrit Sans. i shows varied forms in Philippine(s) Side by side with Sulu gaiah, which shows an unchanged phoneme quality from Sanskrit gaia (Malay gajah, Javanese gaia) are Tagalog & Bisaya gadya, Iloko gadia, showing dy and di equivalents of i. Sulu jaga retains the quality of the phoneme as Malay jaga <Sanskrit jagr (but, ef. Tagalog daga (T)). Sanskrit jala> Tagalog & Pampanga dala :: Mal, jala, Sanskrit jIva > Bikol, Bisaya, South Mangyan & Tagalog diwa :: Malay iiva. It is evident in these examples that Mal, does not influence the development of phoneme in the Philippine(s), except Sulu. It may be deduced

-189 from the above illustrations that j>dy or di when medial, and d when initial. neme If. Bisaya & Tagalog pugay is not native Malayo-Polynesian, it is probable that it is Sanskrit nuja. The i>g-phenomenon, apparently, is an isolated case in the development of the phoBut, of. Sanskrit puja >Tagalog & Pampanga puri and South Mangyan Perhaps, Tagalog (& Pampanga) puri developed from Javanese There must have been an older Tagalog form, e.g., pudi (<Javanese pudii, i being lost on its becoming Tagalog), from which the ?new form, puri, developed. d became r due to the phonetic law in Tagalog, which states that dis lingualized when intervocalic. South Mangyan pudi is not afpudi. pudii. fected by such phonetic law. J in Sanskrit kubia is either lost or assimilated by the preceding labial - Tagalog kuba, Bisaya kubong, South Mangyan kaba, and Iloko kubbo. Sanskrit t. d>Phil, t. d. With no exceptions, Sanskrit cerebrals develop to dentals in Philippine(s) Sanskrit koti > Tagalog kati, "ten millions", Sanskrit kostha > Tagalog kuta, Sulu kota, Iloko, Bisaya & Maranao kota :: Malay kota, Javanese kuta. ghatika > Maranao & Sulu kotika, by the influence of its development in Malay katika. Sanskrit patola >Tagalog patola. Sanskrit As would have been expected, Sanskrit tt in bhattara

-190 should have become t in Tagalog, Bisaya & South Mangyan bathala, by the influence of Malay (or Javanese) t in batara. But, with t appears the metathetic h. Sanskrit pariprsta > South Mangyan & Tagalog alipusta. Sanskrit garuda > Maranao garoda. d>d by the influence of its earlier development in Malay geroda, garuda, Javanese garuda. Malay denda may have influenced Iloko dondon (cf. Toba Batak dandan) <Sanskrit danda, Sanskrit bhanda > Tagalog & Pampanga bandi, Bikol & Bisaya bahandi, Sanskrit mandala > Tagalog mandala, etc. Sanskrit t. d >Philippine(s), in various forms. Javanese utara is the form that influences Sulu utara, like Bisaya utala Sanskrit uttara. It shows that one of the t's is either assimilated or dropped in the process of borrowing. Sanskrit > Philippine(s) s: Sanskrit tila > Tagalog & Pampanga sulasi. The development in Philippine(s) is due to the influence of the intermediate Malay selasih (Javanese selasih), beside telasi (vide §2.25, p. 175). forms, The Sanskrit conjunct, d-dh, generally drops the unaspirated phoneme. Sanskrit buddhi > Tagalog & Bisaya budhi, var. Of., however, Malay budi, which loses the aspirate. An analogous development occurs to Sanskrit siddhi and siddha > Malay sidi, sida, but in Tagalog, the d is lost as in the burhi.

-191 former word - sidhi, si dha. The conjunct appears to be assimilated by the preceding lingual in Tagalog maharlika, which later on develops into a liquid. This is due to the principle that no two linguals are allowed to be contiguous in Tagalog The intermediate forms, however, retain the rd conjunct Malay merdeheka, merdeka. Sanskrit d in pada is completely lost in Philippine(s), and seems to have been replaced by an intervocalic pausa Bisaya, South Mangyan & Tagalog paa, MNeg. pa'a. Cf. Malay pada. is Sans, in origin - W. E. Maxwell 49 writes that the Malay title, magat, magadha, "the son of a vaisya by a ksatriya woman". R. O. Winstedt attests the derivation and writes megat, "the title of the chief who is a noble on one side", originally signifying "descent from one of lesser rank married to a princess". It is certain that Tagalog magat, "a title of a petty chief in pre-Spanish Philippines", is Sanskrit in origin, with Malay as the intervening form. Graphically, Sanskrit dh Malay & Tagalog 1, with syncopation of the final vowel. Sanskrit p. b> Philippine(s) p. b, p. Sanskrit p does not change in Philippine(s) except in one doubtful case, e.g., Bisaya banyaga, 49"Two Malay Myths: The Princess of the Foam, and the Raja of the Bamboo", Journal of the Royal Asiatic Society, xiii, p. 506, fn. 1.

-192- "mean, base" <Sanskrit pandaka, "eunuch, weakling". Since the phoneme does not have another development than in the above example, only a few illustrations may be cited: Sanskrit paduka > Tagalog paruka, Iloko palloka :: Malay paduka, Sanskrit paryanka, palyanka >Iloko palangka :: Malay pelankin, pelanki. Sanskrit b > Philippine(s) b. E.g., Sanskrit bala>Tagalog bala, South Mangyan balam, Sanskrit blia > Sulu biji :: Malay biji, Sanskrit buddhiman > Maranao bodiman (Bisaya & Tagalog budhi < Sanskrit buddhi) :: budiman. These are a few examples of the unchanged quality of Sanskrit b > Philippine(s) b. But, in one instance, it is found to develop into its equivalent tenuis - Sanskrit bana >Philippine(s) pana (Sulu panah :: Malay panah). 50 The devSans. r. 1>Philippine(s) r. 1, and vice versa. elopment of the Sanskrit lingual and liquid is quite constant in Philippine(s), except in a few cases where a d or 1 is found to be 1 or 1. Apparently, the intervening languages do not influence the development of these two phonemes in the Philippine(s) languages. First group - Constant Sanskrit r> Philippine(s) 1: Sanskrit yantra > Tagalog gantala :: Malay jentera, gentala, Sanskrit tara > Bik, & Tagalog tala, Iloko (kamonta)tala :: Old Javanese tara, Sanskrit 50 cr. the Principle of Rhotacism in Vedic language.

-193 mantra > Bikol, Pampanga & Tagalog mantala :: Malay mantala :: Malay mantera, Sanskrit cakra Bisaya & Tagalog sakla :: Malay chakera, Sanskrit samanantara >Tagalog samantala :: Malay sementera, Sanskrit sighra > Tagalog sigla :: Malay sigera. The liquid development of the Sanskrit trill is found to be inconstant, where one or two of the Philippine(s) retains the Sanskrit lingual, by the influence of Malay Maranao has arga, beside Tagalog, Bisaya, & South Mangyan halaga, Sulu halga :: Malay harga <Sanskrit argha. A var. of Sulu sutra, beside sutla (South Mangyan & Tagalog sutla, Bisaya sukla) retains the lingual Malay sutera), Sanskrit svara Maranao suara, Sulu suala :: Malay suara, Sanskrit carita > Iloko sari ta, South Mangyan & Tagalog salita :: Malay cherita, Sanskrit marica >Maranao marisa, Bisaya malisa :: Malay mericha. Sanskrit vrata (Atapas) > Bisaya, Pampanga & Tagalog balata (Malay baratapa), Sanskrit putri > Bisaya putli, Maranao potri :: Malay puteri, Javanese putri, Sanskrit uttara > Sulu utara, Bisaya utala, etc. It will be noted that Tagalog and apparently Bisaya maintain a consistent 1 (Maranao maintains a constant I (Bisaya and Tagalog draws comment. This may be explained in the context of the RLD law in the Philippine(s) languages and

-194- 51 consequently in the Malayo-Polynesian languages." The law may be stated briefly - that where r exists in one language or dialect, it is 1 or d in the other languages or dialects. But, it may appear that 1 in this context can not be applied to the loan-words, particularly Sanskrit , for there is only the F-1 complex, ₫ being ultimately traceable to the original Malayo-Polynesian 52 Note, however, an isolated instance where a Sanskrit loan-word figures within the law Sanskrit paduka > Malay paduka :: Tagalog paruka :: Iloko palloka. Three Malayo-Polynesian languages are involved, one of which is nonPhil. While Bisaya lipaka Sanskrit dipaka or probably paripaka shows a 1-d or 1-r correspondence, there are no other examples of this kind. 2.29. OTHER PHONETIC CHANGES 2.291. Haplology. This process of change of Sanskrit loan-words in the Philippine(s) may have its beginnings in the intervening languages. Classic examples of this process 51 Carlos Everett Conant, "The RGH Law in Philippine Languages", Journal of the American Oriental Society, xxxi, 1911, p. 83-85 (The RLD Law Series). C. E. Conant "The Pepet Law in the Philippine Languages", Filipiniana, II Linguistics, p. 126-128. R. Brandstetter (op. cit., p. 132-133, 267-268) earlier treats this phonetphenomenon. 52 Dyen, op. cit., p. 227-228.

-195 are: Old Javanese pratyantara, "next in succession" <Sanskrit pratyanantara, "id.", Samantara <Sanskrit samanantara, "shortly after", etc.53 The second Old Javanese example is the ancestor of Tagalog samantala, "meanwhile, interval", and Malay sementera. The loss or dropping of one of the identical syllables in the two following words may be due to the svarabhakti vowel that appears Sanskrit sampratyaya > Tagalog sampalataya, and Sanskrit compound ?vrata tapas >?Maranao baratapa. kali may be the shortened form of Sanskrit sahakala. The first member of the compound drops the second syllable, ha, owing perhaps to the seemingly identical first syllable of the second member of the compound, hence, sakala (> ?South Mangyan sakali :: Mal, sakali). 2.292. Syncope. Tagalog saThe initial vowel syncopation appears to be more common than in any other position, e.g., medial or final. To cite a few examples: Sanskrit abhyasa > South Mangyan & Tagalog bihasa, Sulu biaksa :: Malay biasa, biyasa, Sanskrit ayuta >Tagalog yuta, Maranao juta :: Malay yuta, diuta, Sanskrit alabu >Sulu labu :: Malay labu, Sanskrit asuya > Iloko & Tagalog suva, Iloko dava <Sanskrit udaya (> Lit. Javanese udaya). In the final vowel loss, few examples may be cited Sanskrit vikala > Tagalog bigal, Sanskrit samya > Bisaya sami, Sanskrit 53 Gonda, p. 237.

-196 capeta > Iloko sipat, Sanskrit athava > Malay atau > Sulu atau, ataw. Middle vowel loss is evident in Sulu drahka, a deviation from the common development in Malay derhaka (Javanese duraka) <Sanskrit drohaka. ping. 2.293. Consonantal loss and syllabic or word dropConsonantal loss has been partly noticed in one or two of the phonetic changes discussed above. tion, two examples may be cited In addiSans. cukra > Philippine(s) suka, soka (beside Sulu sukak', with the inexplicable hamzah). If Bisaya yugi is derived from Sanskrit yogin, and not from the nom. yogi, a consonantal loss is evident. Syllabic loss is apparent in Bikol, Bisaya, Tagalog & South Mangyan mahal, if it is Sanskrit maharha (cf. Tavera's derivation from maha-). Syllabic loss is also apparent in the development of Iloko bangsal (<Malay bangsal Sanskrit Vaniiyasala), which evinces a syncopated final a. The first part of the Sanskrit ? compound undergoes considerable change in Bengali. Iloko tanikala (Bisaya & Tagalog talikala) shows a peculiar development if Sanskrit tanka (gold) srnkhala are its ancestors. If so, its development may have proceeded in this manner first, the syllable srn is lost, thus tanka-

-197 khala; then by haplology, one of the identical syllables is lost, thus tankala; and lastly, the epenthetic i must have appeared in analogy with Bisaya & Tagalog tali°. Loss of one member of the compound is also known, though it is not common. Only one of these instances may stand scrutiny. The second member of the Sanskrit compound cudamani is lost in Tagalog & Bisaya, thus there is only su- 1 a. Cf. this, however, with Sanskrit sila. Iloko & Tagalog upa does not necessarily show a developThe ment in the manner similar to the preceding word. form is indeed difficult to recognize without the assumed intermediate form. Sanskrit utpatti (< ut / /pat), "production," undergoes considerable change in Malay and Javanese upeti, Javanese var. upekti, "tribute". If Iloko & Tagalog upa is a loan-word from the remote Sanskrit , it is certainly developed from Malay upeti, with a final syllabic loss, and the pepet vowel showing its a-correspondence in Iloko & Tagalog Cf., however, the various Sans, forms in which upa may have been originally used upakara, "help, any financial help"; upada, "present"; upahara, "offering something worthy". The Philippine(s) upa may have survived all the forms but still retaining the fundamental sense, in the course of the word's travel from Sanskrit - -

-198- 2.294. Vocalic and consonantal gemination. With no apparent sign of external influence, Central Cagayan Negrito mukaat geminates the final a <Sanskrit mukha :: Tagalog mukha, Javanese mukha. Note, however, the final plosive. Maranao maana and South Mangyan maan <Sanskrit mana. It appears that South Mangyan transposes the final a. Gemination of the consonant is apparent in Philippine(s) aparri, if Sanskrit upari is its ancestor. 2.295. Some sporadic changes. Changes which do As it is expected, the not have analogous formations in Malay (or Javanese) can not be ignored because they show independence from the phonetic laws of the Island languages. development of Sanskrit u or ur in murcha to Sulu oe in moeka would have followed the normal development of Sanskrit phonemes as seen in Malay murka, etc. Similarly, Sanskrit a or ra in pranasa > Bisaya u in punasa is unexpected. This The precipitate development of Sanskrit sv in svami to g in Bisaya gami has been discussed partly in §2.217. development is rather puzzling. If Iloko pardaya is not native Philippine(s) or Malayo-Polynesian, it is Sanskrit parivada, with all the phonetic possibilities in explaining the change of the original to Iloko Cf. Sanskrit rahasya >Tagalog rahuyo.

-199 Where haplology is expected, Iloko kamalala < Sanskrit kamalola, shows an adoption of the original except with Sanskrit o > Iloko a, by the influence of the neighbouring a's. Apparently, Sanskrit gala becomes reduplicated in Iloko & Tagalog Cf. Malay gegala, which galagala, Malay & Javanese galagala. reduplicates the initial syllable. a non-reduplicated form. But, Maranao gala shows Whether Bisaya & Tagalog la (Malay le) in lagundi is a development from Sanskrit nir- in nirgundi or not, it is worth being given some attention here. The e-a correspondence in Malay<>Bisaya & Tagalog has been explained above. The prefix le or la 54 might have been affixed to the dissyllabic development of the Sanskrit in Malay: gundi, owing to the tendency of dissyllabism in Malayo-Polynesian languages. Thus: -?Malay le or?Bisaya & Tagalog 1 a / Sanskrit gundi. 2.296. Analogous phonetic formations. This phonetic formation may also be rightly called onomatopoeic analogy. A few of the words that fall under this development are kinship terms. Sanskrit ama "father" :: Philippine(s) ama, "id."; Sanskrit tata, "father" :: Philippine(s) tata, "uncle, father", var. Iloko tatang, Bisaya & Tagalog tatay; Sanskrit nana, 54 J. Pijnappel, "On the Roots of the Malay Language' Journal of the Royal Asiatic Society (Straits Branch), xvi, p. 256.

-200- "mother" :: Philippine(s) nana, "aunt, mother", var. Iloko nanang, Tagalog nanay; Sanskrit mama, "uncle, only in vocative singular as term for affection among animals in fables" :: Tagalog ma "uncle, an address to an elderly man" :: Malay mama or mamak, "uncle, a title of respect."55 55 H. W. Hamilton, "Hindustani, Tamil, Sanskrit and other Loan-words in Malay", Journal of the Royal Asiatic Society (Straits Branch), 1 xxx. Mr. Hamilton derives the Malay nama or mamak, from Sanskrit mama, "uncle, title of respect", but it is more likely that it is Sanskrit mama.