*w   

*wáak crow

Note:   Forms such as Hanunóo wák suggest that this reconstruction is identical to *uak ‘crow’. Bontok wáɁak, however, appears to be incompatible with this interpretation.

*wada be, exist, have; wealthy; not exist, not have

Note:   Also Isneg wadʔóy ‘there is (used in songs)’, Bontok wadʔáy ‘bring about; make; form; create; father a child; there is, there exists’, Ifugaw wa ‘there is’, Ilokano addá ‘to have; to be (somewhere), there is, there are, there was, there were; some’. Dempwolff (1934-38) reconstructed *wada ‘Vorhandensein, Nichtsein’. This apparently contradictory gloss is difficult to improve: the idea of existence or possession is clearly reflected in the northern Philippines, in Iban, Malay, Mentawai, and Chamorro, while a negative sense is clearly reflected in the central and southern Philippines, in Maloh, in Old Javanese and Modern Javanese, and in UMAT of central Sulawesi. These opposing senses do not coexist (without some morphological or lexical modification) in any attested language, and it is unclear how they can be reconciled in a reconstructed gloss. Forms such as Tagalog waláʔ and Cebuano waráy appear to be morphologically complex, and show a puzzling parallelism to the morphology of vocatives (Blust 1979).

*-wadaŋ collarbone, clavicle

Note:   Also Kankanaey ali-madáŋ ‘clavicle, collarbone’. Although this morpheme took the *qali-, *kali- prefix, the particular prefixal variant is presently indeterminate, as Puyuma reflects *qaNi-, Itbayaten reflects *qari-, the Cordilleran languages reflect *qali-, and Toba Batak reflects *kali-.

*wadwad to distribute, spead out evenly

*wagwág₁ rice variety

*wagwág₂ shake something vigorously (as a sack to

Note:   Also Bontok wayagwag ‘shake something, as clothes in order to remove ants’.

*wahiR fresh water; stream, river

Note:   Also Ifugaw waél ‘creek’, Iban aiʔ ‘water, liquid, liquor, juice, sap’, Sasak aiʔ ‘water’, Manggarai wae ‘water’, Rembong waeʔ ‘water; river; liquid’, Ngadha vae ‘liquid, juice, water, brook, lake’, Rotinese oe ‘water, liquid, sap, juice, semen’, Tetun we ‘water’, we-n ‘juice, sap’, Paulohi waele ‘water’, Kamarian waer ‘water’, Buruese wae ‘water, stream, creek, river’, wae-n ‘fluid’. PMP *wahiR, one of the most widespread and characteristic Austronesian morphemes, often has reflexes in place names.

*wai mango sp.

Note:   Also Maranao wani ‘Mangifera odorata Griff.’. Chowning (1963) reconstructs "PMel" *wai ‘Mangifera indica’. Since PMP *pahuq evidently referred to the Mangifera indica, I assume that PMP *wai designated other varieties, although *wai may well have become a generic term in POc. Tsuchida (1977) attributes Tsou suai, Bunun (Ishbukun) suaiʔ ‘mango’ to borrowing from Southern Minangkabau soai-a. If the direction of borrowing was the reverse of that assumed by Tsuchida, however, these forms, in conjunction with the above cognate set, may indicate a PAn etymon *Suai.

*waip water container

Note:   The OC forms cited here may reflect POc *waiR ‘fresh water’. If so, all that can be inferred is a second gloss for this reconstructed form.

*waka knob on the head of a club

*wákaŋ to walk with large strides

Note:   With root *-kaŋ₁ ‘spread apart, as the legs’.

*wakaq split

Note:   With root *-kaq₂ ‘open forcibly, split’.

*wakaR root

Note:   Also Bintulu waka ‘root’ (Ray 1913). Reflexes of *wakaR ‘root’ and *wakat ‘mangrove root’ may be difficult to distinguish in OC languages (which typically have lost original final consonants). However, close attention to the semantics suggests that only *wakaR survived in POc. This conclusion is strengthened by reflexes of *wakaR in Oceanic languages which preserve final consonants (as Mussau), and by the apparently suffixed forms with *-a and *-i (although the morphology of these longer forms remains obscure). Despite its reflexes in many WMP languages, PMP *wakaR evidently did not mean ‘vine’ (cf. *waRej ‘vine’), although in addition to ‘root (generic)’ it may have referred more particularly to creeping roots.

*wakas₁ loosen, undo, uncover

Note:   With root *-kas₂ ‘loosen, undo, untie’.

*wakas₂ to finish, to end

*wakat mangrove root

Note:   Also Seru ukat ‘root’. Although no CMP language apart from Atoni is known to reflect *wakat in the meaning ‘root’, it seems clear from Bikol, Aklanon, Cebuano wákat and from Mongondow wakat-an that we are dealing here with a single cognate set. Since PMP *wakaR evidently meant ‘root’ (generic) and PMP *teŋeR meant ‘mangrove’ the most likely meaning of *wakat is ‘mangrove roots’. Dempwolff's (1934-38) attempt to relate Malagasy vahatra ‘a colony in progress of settlement’ to *wakaR appears to be misguided, but the Malagasy form may be a metaphorical development from *wakat.

*wákat scatter, strew about

*waket mangrove root (?)

*wakir kind of root

*waksí shake off, flick off

Note:   Also Casiguran Dumagat wagseh ‘push or shake something off you; flick or throw something off to the side’.

*wakwak bird sp.; to caw, crow

Note:   Also Puyuma (Tamalakaw) waHwaH ‘crow’. This form appears to be onomatopoetic, and may ultimately be a reduplication of *uak ‘crow’.

*wali₁ to paint; smear or rub on

*wali₂ to lecture someone

*walís to sweep; broom

Note:   Possibly a Tagalog loan distribution.

*walu eight

Note:   I interpret the agreement of the following morphologically complex reflexes of *walu as products of convergence: 1. Paiwan kaR-waru waru ‘eight each’, Bikol kag-waló ‘eighteen’, 2. Bikol ka-waló-an ‘eighths’, Madurese ka-ballu-an ‘group of eight’.

*walwál work (object, instrument) from side to side

*walwáR work (object, instrument) from side to side

*wanaN right (side, hand, direction)

*wane wane straight, direct; flat, level

Note:   PEAdm *mʷanen (Pak mʷanen, Loniu mʷenen, Nali monen, Penchal mʷanen(-en), Ahus monen-en ‘straight’) may be related if the following assumptions are justified: (1) POc *wane wane sometimes took the attributive suffix *-na; (2) In PEADM the syncopation of a vowel in the environment VC__CV, which is attested in some other forms, produced *wanwane-na; (3) The derivative cluster *nw assimilated to *mw, yielding *wamwane-na; (4) The initial syllable was then lost. POc final vowels were regularly lost in all languages of the Eastern Admiralties.

*wani honeybee

Note:   Also Iban mañi ‘bees’, Jarai heni ‘honeybee’, Makassarese bani ‘bee’, Sika wane ‘honeybee’. PMP *ñ, which is suggested by the Melanau (Mukah), Iban, Singhi Land Dayak and Maloh reflexes, is contradicted by Jarai, Bare'e, Proto-South Sulawesi, and Hawu. I assume that the Bornean witnesses show secondary palatalization of *n. Mills (1975:883) posits "Proto-Indonesian" ? *wani ‘bee’. The prototype of this comparison may have been *qawani.

This form evidently referred to the ordinary honeybee as opposed to the smaller Apis indica (*qari-ñuan, *qani-Ruan), or the stingless Trigona spp., and in this function it replaced PAn *waNu ‘honeybee’.

*waNan right (side, hand, direction)

Note:   Also Saisiyat kaʔnal, Ma'anyan kawan, Uma kaʔana ‘right side’. PAn *ka- may have distinguished right/left hand from right/left side/direction, but the function of this minimally productive affix remains unclear. I take the similarity of the longer affixes in Gorontalo olo-wala, Mongondow kolo-anan, and Soboyo kala-wanañ ‘right side’ to be a product of convergence.

Unlike PAn *wiRi ‘left side’, there is little evidence that PAn *wanaN, waNan was prefixed with *ma- in POc. Ali mawaŋ ‘right side’ may indicate that the prefixed form occurred, but the reflexes in Manam, Wogeo, and some other languages show clearly that *wanan was found in POc, whereas *wiRi was not. Finally, at least three widely separated languages reflect *nawaN rather than *wanaN. I take this to be the result of independent metatheses rather than as evidence for a doublet.

*waŋáwaŋ wide open space

Note:   With root *-waŋ ‘wide open space’.

*waŋis bare the teeth

Note:   Also Kanakanabu anísi, Tsou hisi, Proto-Rukai *valisi, Puyuma (Tamalakaw) wali, Amis wadis ‘tooth, teeth’, Bunun vani ‘boar’; Cebuano waŋsiʔ ‘having one or both of the lips turned up so that it is impossible to close the mouth’. With root *-ŋis ‘bare the teeth’.

*waŋka canoe

Note:   Based on the comparison Javanese waŋkaŋ ‘ship’, Ngaju Dayak waŋkaŋ ‘Chinese junk’, Fijian waŋga ‘canoe’, Tongan, Futunan vaka, Samoan vaʔa ‘canoe’ Dempwolff (1934-38) reconstructed *waŋkaŋ ‘water vehicle, ship, canoe’. This long accepted etymology actually is highly problematic.

First, reflexes with a final nasal are known in only five languages, and in each case these refer specifically to Chinese junks: Ngaju Dayak waŋkaŋ ‘Chinese junk’, tanah waŋkaŋ ‘China’; Malay waŋkaŋ ‘Chinese ocean-going junk’, limau waŋkaŋ ‘imported Chinese oranges’; Acehnese, Javanese, Makassarese waŋkaŋ ‘Chinese junk (boat)’.

Second, neither the Malay nor the Javanese words can regularly reflect *waŋkaŋ, since *w- disappeared in Malay and *wa- became Javanese o.

Moreover, the word in Ngaju Dayak, which is spoken in the interior river systems of Southeast Borneo, clearly is a loan from the coastal Banjarese. It appears likely, then, that waŋkaŋ is a loan from some southern form of Chinese (Fukienese? Hakka?) which was borrowed into Malay and Javanese.

In apparent contradiction to this conclusion are reflexes in CMP, South Halmahera-West New Guinea, and especially OC languages, where the probability of early borrowing from Chinese is all but non-existent. Since most OC languages have lost original final consonants, many OC reflexes could derive from *waŋkaŋ. In those languages which normally preserve the final consonant, however, no final consonant is found in this word (Dobuan, Roviana), thus supporting *waŋka, not *waŋkaŋ. Doran (1981) has chosen this form as the title of a book on Austronesian canoe origins, but he combines a genuine Oceanic cognate set with such non-cognate WMP forms as Tagalog baŋkaʔ ‘boat’, Malagasy lakana (Prov. laka) ‘canoe’.

It is now simplest to conclude that *waŋka was an innovation which appeared in northeastern Indonesia prior to the separation of the CMP and EMP language groups from one another, but after the separation of PCEMP from all other Austronesian languages. Under this interpretation, the similarity of the native term *waŋka to the apparent loanword waŋkaŋ must be considered a remarkable coincidence. Under this interpretation, Mongondow waŋga (for expected **waŋka) ‘boat’ is best regarded as a loan from some still undetermined CMP source.

*waŋkaŋ spread the legs apart

Note:   With root *-kaŋ₁ ‘spread apart, as the legs’. Amis wakaŋ ‘lie on one's back’ may be related.

*waŋwáŋ gaping expanse

Note:   Also Bikol waŋíwaŋ ‘gaping (a hole)’. With root *-waŋ ‘wide open space’. Bontok waŋwáŋ ‘river’ and similar forms in other languages of northern Luzon may also be related (cf. the semantic range of reflexes of *bawaŋ ‘open expanse of land or water’).

*wao forest, uninhabited land

Note:   Also Proto-Micronesian *walu ‘forest, woods, bushy area’.

*waqay foot, leg

Note:   Also Manggarai waʔi ‘leg, lower part of something’, Rembong waʔi ‘leg’, Ngadha vai ‘foot, leg; footprint; rung of a ladder; fencepost; lath’, Sika waʔi(-ŋ) ‘foot’, Mambai oi-n ‘foot, leg’, Gedaged wae-n ‘the arms of an octopus’. Palauan oach and CMP forms such as Manggarai waʔi may reflect a doublet *waqi.