pain

Dempwolff (1938) posited *kedut ‘pain’, but theTagalog form is irregular (expected **kurót), and is a not a convincing semantic match forJavanesekedut.

palate

palm tree sp.

palm:   sago palm

The resemblance ofBalineseental to the other forms cited here is attributed to chance. TheLau andArosi forms must reflect earlier *osa, andRotumanota must be a loan, although its source remains problematic. This leavesMotaota,Anutaota 'sago palm' as the only clearly cognate forms, but given the evidence for borrowing inRotuman this distribution is perhaps best regarded as a product of diffusion.

Palolo worm

pant, be out of breath

Tempting as it is to accept this as a valid comparison, the sound correspondences simply do not work. ThePuyuma vowels disagree with those of thePaiwan andIbaloy forms, and although the latter two languages appear to show a perfect agreement,Paiwan /s/ reflects *S andIbaloy /s/ reflects *s, making reconstruction impossible. Like many other bases with initial velar nasal, this one probably is a product of convergent innovation involving the phonesthemeŋ-, widely associated in Austronesian languages with the nasal-oral area (Blust 2003a).

pardon, forgiveness

The forms inJavanese,Sundanese andBalinese are cognate, but the similarity of these toTagalogapulaʔ is due to chance.

pare:   to peel, pare

parent-in-law

part fraction

particle:   affirmative particle

particle:   clause-final particle

particle

TheTagalog linker and theHanunóo conjunctive particle are assumed to be cognate. The resemblance of the other forms to these is attributed to chance.

passive affix

The suffixed *-a inOceanic languages probably is identical to *-a '3sg. object'. The prefixeda- inMalagasy andFijian is assumed to have no historical connection, either between these two languages, or with any of the suffixes cited here.

past tense marker

Apparently reconstructible forPCEP. ThePaiwan andAmis forms disagree in pointing to *Si and *i respectively. I take the resemblance ofPCEP *i to all other forms cited here as a product of chance.

paste

Most likely a chance resemblance.Dempwolff (1938) reconstructed Uraustronesisch *lepa ‘paste’.

pattern:   textile pattern

paunch

pearl

peel:   to peel, pare

peel:   to peel

Reconstructible asPCP *isi. The resemblance ofRotineseisi toPCP *isi is attributed to chance.

peel off, separate

Chance, although both forms appear to contain the root *-bak₃ 'split off, separate'.

peeling, rind

Chance. For an alternative etymological connection forCebuanopaflut cf.POc *paro. For an alternative view of theCebuano:Malay comparison which is now rejected cf.Blust (1970).

penetrate:   pierce, penetrate

Dempwolff (1938) posited *serep ‘penetrate’, a form that was accepted inBlust (1972), but for which the supporting evidence is now considered unconvincing.

perform:   do, perform

permission

Based on theTagalog andMalagasy formsDempwolff (1938) proposed ‘Uraustronesisch’ *tulut ‘permission’, despite a lack of comparative support for his proposed gloss.

persuade:   entice, persuade

The resemblance ofKankanaeyimók to theCentral Philippine forms is attributed to chance.

pick up

Mills (1981:75), who cites both theBuginese andMakassarese words with a simplexk, proposedPAn *si(ŋ)kit ‘to pick up’, but the known evidence does not appear strong enough to rule out chance convergence.

piece broken off:   fragment, piece broken off

pierce, skewer

Dempwolff (1938) included these forms under ‘Uraustronesisch’ *zuluk ‘pierce, penetrate’ (Stechen). However, there is little to recommend this comparison, which appears to be a collection of formally compatible words that have almost nothing else in common. It is true that the meaning ‘penetrate, bore through’ inToba Batakman-julluk is fairly similar to ‘put in, insert, sheathe’ inSamoansulu, but the medial geminate is unexplained (Dempwolff wroteju<l>luk), and it is further noteworthy thatTonganhulu ‘to fix one’s loincloth by tucking in one end of it at the waist; to tidy up the edges of a mat by tucking in the protruding ends’, was omitted despite its obvious cognation, presumably because it would weaken the argument for comparing theSamoan form with those in insular Southeast Asia

pierce, penetrate

Dempwolff (1938) posited *serep ‘penetrate’, a form that was accepted inBlust (1972), but for which the supporting evidence is now considered unconvincing.

pig

pigpen

pinch

pith

pity, compassion

The similarity ofMakassareseballasaʔ to the other forms cited here probably is due to chance. The remainder of this comparison is best regarded as a late innovation in western Indonesia, with likely borrowing either fromMalay or fromJavanese.

place:   put in place

place of burning

Dempwolff (1938) proposed *(s)eRa(q) ‘place of burning’, but this comparison is both formally and semantically problematic, and is best treated as a product of chance convergence.

place of residence

Despite the clear phonological irregularity in the initial consonant, and the divergent semantics of theMalay and Polynesian formsDempwolff (1938) positedPAn *tempet ‘place of residence’.

plan, design, project

plank:   board, plank

Dempwolff reconstructed *bani(r), andTsuchida (1976) posits *baNiR/. I regardSaaroa,Tsou,Paiwan andNgaju Dayak,Malay,Acehnese,Sundanese,Sasak as forming two geographically restricted cognate sets. The resemblance of these sets to one another and toSelaru andYapese is attributed to chance.

plant:   food plant

plant sp.

plant sp.

plant sp.

plant sp.

plant sp.

plant sp.

plant sp.

plant sp.

plant sp.

Chance, or if related,Aklanon,Cebuano,Maranao,Mongondow reflect an innovation inPCPh.

plant sp.

Chance.Cebuanopálaw 'large vigorously growing aroid:Cyrtosperma/ merkushii',Maranaopalao 'cultivated taro:Xanthosoma violaceum Schott. suggest aPCPh *pálaw 'large aroid:Cyrtosperma spp.'. TheTagalog word shows secondary suffixation.

plant sp.

plant sp.

TheOld Javanese andBalinese words are cognate. The similarity of theIsneg andKayan forms to these and to each other is due to chance.

plant sp.

plant sp.

plant shoot:   sprout, plant, shoot

Dempwolff (1938) citedFijiani-so-somi ‘seedling’, but this is not the sense of this term inBauan Fijian as given byCapell (1968), or inWayan as given byPawley and Sayaba (2003). In the absence of further evidence this comparison is best treated as a product of chance.

plants:   covering membrane of plants

pledge:   vow, pledge

This comparison proposed byDempwolff (1938) appears to bring together completely unrelated forms. I am unable to findTagalogsahot in eitherPanganiban (1966) orEnglish (1986), andSamoansau does not occur in this meaning in any dictionary I have been able to consult, includingPratt (1893), which was Dempwolff’s source.

plot evil

pluck

pluck (as fruit)

plural marker

point:   sharpen to a point

Chance?

point, tip

Dempwolff proposed *zuru ‘point, tip’ (dbl. *duRu ‘corner’), and included such semantically divergent forms asMalayjuru ‘trained worker (at some occupation other than a handicraft)’ andJavanesejuru ‘person who performs a certain [skilled] job’, under the general gloss ‘master’ (also cf.Old Javanesejuru ‘head, leader, chief (of a division, military or administrative; of a group or trade); tradesman, trained worker’). The entire collection seems extremely forced, and is best abandoned as ill-conceived.

point out directions

Dempwolff (1938) reconstructed ‘Uraustronesisch *taki ‘point out directions’.

pole, mast

TheNgaju Dayak,Malay andJavanese forms are related, but probably a product of borrowing fromMalay (withNgaju Dayaktihaŋ fromBanjarese). The similarity of the Oceanic forms to these is best treated as a product of chance.Dempwolff (1938) reconstructed ‘Uraustronesisch’ *tiaŋ ‘pole, rod’ (Pfahl, Stange) and dismissed theNgaju Dayak form as irregular, butOld Javanese confirms the medial consonant (< *q).

poles:   suspend from poles

poor

pouch:   basket pouch

Dempwolff (1938) posited *kanTuŋ ‘Korb, Tasche am Kleid’ (basket, pouch on the clothing). However, as he noted, the last vowel of the Oceanic forms fails to agree with that of his reconstruction. The forms in western Indonesia appear to be loans fromMalay which theFijian,Tongan andSamoan forms resemble by chance.

pound:   beat, pound

pound:   beat, pound

Dempwolff (1938) used this comparison withJavanesekəwut ‘beat out’ instead of theSundanese form cited here. However, I am unable to find aJavanese form meeting this description in eitherHorne (1974) orPigeaud (1938), and theTagalog form is both phonologically irregular and semantically deviant, leaving little comparative basis for a reconstruction.

pour:   stream, pour

poverty

Dempwolff (1934-1938) proposed ‘Uraustronesisch’ *dahup ‘weak’.

prayer

TheMalay andJavanese words given here are clearly Arabic loans. TheTagalog word looks like it could be a borrowing fromMalay, but if so it would be the only known case in which aMalayr has been borrowed as aTagalogg, mimicking the correspondence reflectingPAn *R. For now it seems best to consider the resemblance of theTagalog word to those inMalay andJavanese a product of chance.

precipitous:   aslant, steep, precipitous

Dempwolff (1938) proposed *si(dD)aŋ ‘steep, precipitous’, but this comparison appears to be confined toNgaju Dayak andMalagasy. It is thus best treated as ‘noise’ not on grounds of false cognation, but rather on the grounds that it cannot be reconstructed for a proto-language of any considerable time-depth.

prefix

Despite a problem with the medial consonant correspondence, theBikol,Aklanon, andHiligaynon affixes appear to be cognate. The similarity of these forms toKayanpala-,RALUwara- is attributed to chance.

pregnant

prepared:   ready, prepared

Based on forms inToba Batak andJavaneseDempwolff (1938) posited *sikep ‘ready, prepared’. However, theKaro Batak form shows thatProto-Batak had *siŋkeb, which is phonologically incompatible with the forms inOld Javanese,Javanese,Balinese andSasak, leaving either a comparison with relatively little time-depth, or a loan distribution.

preserved, protected

None of these forms appear to be related, yetDempwolff (1938) combined them under Uraustronesisch *le(m)bu ‘preserved, protected, hedged in’ (Eingehegt), in a proposed etymology that is best dismissed as a product of chance.

press

press down

pretend:   feign, pretend

Dempwolff (1938) proposed *sisip ‘to feign, pretend’ based on data fromTagalog,Toba Batak,Javanese andSamoan. I am unable to find an equivalent of hisTagalog form in eitherPanganiban (1966) orEnglish (1986), or an equivalent of hisSamoan expression in eitherPratt (1893) orMilner (1966). However, even if these words were taken at face value this comparison would remain unconvincing without further support.

pretty

project:   plan, design, project

preserved, protected

None of these forms appear to be related, yetDempwolff (1938) combined them under Uraustronesisch *le(m)bu ‘preserved, protected, hedged in’ (Eingehegt), in a proposed etymology that is best dismissed as a product of chance.

provided that, in spite of

proximal deictic:   this, proximal deictic

pry up

TheHanunóo andGorontalo forms presumably are related. Although both exhibit regular phonological correspondences with the similar words inNgadha andRotinese , this is not the case withManggaraiNuʔa (where a finall is expected). Since theManggarai,Ngadha andRotinese forms almost certainly are cognate with one another (but not withAsiluluhuʔak) this leaves two low-level cognate sets (oneGCPh, the otherLesser Sundas) and a potpourri of chance resemblances.

pry up

pull up weeds

pulpy leaf

Chance. Although this item might be considered a doublet of *palacipaq,Tae'balaʔbaʔ appears to derive from an infixed form of *bejbej (cf.Tae'baʔbaʔ).

pumpkin:   gourd, pumpkin

Malay non-cognate,Ngaju Dayak possibly an early borrowing fromJavanese.

pure

purpose:   intention, purpose

pursue

This comparison was the subject of a tortuous discussion in the earlier literature based on: 1. an approach to linguistic reconstruction which circumvented the normal scientific procedure of triangulation in the justification of inferences, and 2. a failure to distinguish directly inherited from borrowed forms (Dyen 1953:363ff,Hendon 1964:377ff). TheNgaju Dayak,Iban andMalay words clearly reflectPAn *buRaw 'drive off, chase away, expel' with the atypical, but recurrent change *-aw >u. All other forms in island Southeast Asia appear to beMalay loans, as attested by lexical doubletting inKaro Batak,Toba Batak andDairi-Pakpak Batak, where a directly inherited reflex of *buRaw is found next to theMalay loanword, with semantic differences that are consistent with a borrowing hypothesis. Given the relatively complete integration of this loanword into the affixational systems of some of these languages (e.g.Toba Batak, and the appearance ofburu inOld Javanese texts it appears likely that this word was borrowed fairly early in most languages (perhaps during the classic Sriwijaya period, circa 7th century AD). The similarity of theSoutheast Solomonic forms to those in western Indonesia is attributed to chance.

pursue, chase

push aside

push shove

Probably a chance resemblance.Dempwolff (1938) posited ‘Uraustronesisch’ *tunzaŋ ‘push, shove’ (stossen).

put in place

putrid