s   

sago palm

The resemblance of Balinese ental to the other forms cited here is attributed to chance. The Lau and Arosi forms must reflect earlier *osa, and Rotuman ota must be a loan, although its source remains problematic. This leaves Mota ota, Anuta ota 'sago palm' as the only clearly cognate forms, but given the evidence for borrowing in Rotuman this distribution is perhaps best regarded as a product of diffusion.

sand

satiated

Dempwolff (1938) reconstructed *se(m)buq ‘satiated’, but no reconstruction of any great time-depth appears possible from this comparison.

save, not use

say, tell

say, want to

The resemblance of Buli iwa to the forms in CMP languages probably is due to chance. The CMP forms may be related, but it is not clear that a form found only in Moa, Wetan and Kei must derive from a PCMP etymon.

scab:   sore, scab

scatter

scoop up

On the basis of this comparison Dempwolff (1938) posited ‘Uraustronesisch’ *ca(ŋ)kap ‘to scoop up’.

scour:   rub, scour

Chance. Dempwolff (1938) compared these forms to Tagalog ulós 'wound caused by a pointed instrument; act of striking with a speaŕ and Javanese urus 'affair, matter; to handle, take care of; administer a laxative' under a proposed etymon *ujus.

scrape, scour

The resemblance of the Mongondow form to the similar words in eastern Indonesia is attributed to chance.

scrape

scrape, scour

The resemblance of the Mongondow form to the similar words in eastern Indonesia is attributed to chance.

scrape or scratch with the hand

Chance. Dempwolff (1938) posited ‘Uraustronesisch’ *Ta(ŋ)kur ‘scrape or scratch with the hand’ (mit der Hand scharren). In addition to the above forms he included Tagalog taŋkol ‘hand movement’ (Handbewegung), but no such word appears in either Panganiban (1966) or English (1986), and I am unable to find corresponding forms in other Philippine languages.

scratch:   stroke scratch

Dempwolff posited *kaDus ‘stroke, scratch’ based on these two forms, which appear to have no connection with one another.

scrape or scratch with the hand

Chance. Dempwolff (1938) posited ‘Uraustronesisch’ *Ta(ŋ)kur ‘scrape or scratch with the hand’ (mit der Hand scharren). In addition to the above forms he included Tagalog taŋkol ‘hand movement’ (Handbewegung), but no such word appears in either Panganiban (1966) or English (1986), and I am unable to find corresponding forms in other Philippine languages.

scrotum

This rather fanciful etymology was used by Dempwoff (1938) to reconstruct *kunTul ‘scrotum’. However, the Javanese form is probably a Malay loan, and the Sa'a form a chance resemblance arising from the general meaning ‘round object, lump’ (aside from which *kuntul would regularly yield *Ɂuu, not *ɁuɁu). A far better-supported term for this meaning in PMP *laseR ‘scrotum and testicles’.

scrub:   rub, scrub

scrub:   rub, scrub

The similarity of Pangasinan, Kapampangan, Tagalog is attributed to borrowing from Tagalog. The similarity of these forms to Puyuma, Paiwan and to Itbayaten is regarded as a product of chance.

seed

send

separate:   peel off, separate

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

separate:   to separate

separate:   to separate

separation

set fire to

Wolff (2010:812) proposes PAn *dabedab ‘set fire to’ on the basis of these three words. However, the Kelai form is not reported in Blust (1993), or Amster (1995), nor does a similar word appear in the Lun Dayeh dictionary of Ganang, Crain, and Pearson-Rounds (2006). Wolff’s ‘Kelai dadab’ is thus a ghost form that serves no legitimate purpose in this etymology, and the semantic divergence of the Amis and Tagalog items does not instill confidence that the comparison is valid.

set in order

shake

Chance?

shake out

shame

Following a suggestion first made by Adriani (1928:411), Mills (1981) proposed a slightly different version of this comparison, and posited ‘Proto-Indonesian’ *luŋkut ‘ashamed’. However, the phonological correspondences are irregular, and the Tagalog form is semantically distinct from the words in languages of Sulawesi.

shark

This comparison appears to be a product of chance. Despite several problems with the data Dempwolff (1938) reconstructed *kiu ‘shark’.

sharpen to a point

Chance?

shell

shell

Dempwolff (1938) assigned the above words along with Javanese karah ‘metal piece by which a blade is affixed to a handle’, Toba Batak hara ‘turtle’ and Sa'a kara ‘to scrape, to grate’ to *karaq ‘shell’. However, the Javanese and Sa’a forms are semantically distant from the others, the Toba Batak form does not appear in Warneck (1977), and of the remaining forms Tagalog kála and Malay karah are phonologically incompatible. This leaves just the Tagalog and Malagasy forms as potential comparata, and the similarity between them hardly seems to depart from chance.

shell

Dempwolff (1938) assigned the above words along with Javanese karah ‘metal piece by which a blade is affixed to a handle’, Toba Batak hara ‘turtle’ and Sa'a kara ‘to scrape, to grate’ to *karaq ‘shell’. However, the Javanese and Sa'a forms are semantically distant from the others, the Toba Batak form does not appear in Warneck (1977), and of the remaining forms Tagalog kála and Malay karah are phonologically incompatible. This leaves just the Tagalog and Malagasy forms as potential comparata, and the similarity between them hardly seems to depart from chance.

shellfish sp.

shellfish sp.

shelter, conceal

shield

shine, bright

I am unable to find Tagalog dalaŋ in this meaning in either Panganiban (1966) or English (1986), and the Samoan form, which appears to refer to the shininess produced by a greasy surface, appears only in Pratt (1984). Dempwolff (1938) proposed ‘Uraustronesisch’ *daDaŋ ‘to shine, be bright’, but without better evidence this comparison is best considered a product of chance.

shoot up

shove:   push shove

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

show the teeth

shrill whistle or chirrup

shrub sp.

Dempwolff (1938) cited Tagalog tábal as ‘name of a shrub’, but I cannot find it in any modern dictionary of the language, or in Madulid (2001), and this comparison is consequently best considered a chance resemblance.

shut:   cover, shut

The Bornean forms point to *ateb; the similarity of Tagalog atáb/ and Muna ghonto to these is attributed to chance.

shy

Dempwolff (1934-1938) proposed this comparison (less the Iban form), assuming that the Fijian form shows an otherwise unpreceded syllable metathesis. However, I can find luma in no Fijian dictionary at my disposal, and the form that Dempwolff apparently mean is druma ‘foolish, clumsy, stupid’, a word that has no obvious relationship to those in western Indonesia (most of which may be Malay loans).

sickly

side:   slope, mountain side

side, flank

Dempwolff (1938) proposed PAn *tambaŋ ‘side, opposite side, cross side’, glossing the Malay and Javanese forms as ‘überqueren’ (‘cross over’), and the Fijian form as ‘side’, but this meaning does not occur in Capell (1968). The entire comparison is extremely forced, and best treated as a collection of random similarities.

sing, song

Chance. Dempwolff (1934-38) compared these forms, but was forced to recognize a fossilized affix in the Malagasy word (h-ira), Tagalog and Bikol híla appear to be connected with híla 'pull, drag, tow, haul; a load, as a unit measure for haulage'.

sister-in-law:   brother-in-law, sister-in-law

skewer:   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’ in Toba Batak man-julluk is fairly similar to ‘put in, insert, sheathe’ in Samoan sulu, but the medial geminate is unexplained (Dempwolff wrote ju<l>luk), and it is further noteworthy that Tongan hulu ‘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 the Samoan form with those in insular Southeast Asia

skin disease

Malay bodok is assumed to be a loan from Javanese. The similarity of the Philippine forms to those in western Indonesia is attributed to chance.

skull, cranium

Although initially tempting, this comparison is problematic in two respects: 1. the final correspondence appears to be without parallel, 2. the referent in Philippine languages evidently is the bony structure of the body in general rather than the skull in particular. With regard to the first of these points Zorc (1982:115) has argued for the reconstruction of PAn *-q, *-ʔ, and *-h. Amis foŋoh points to an etymon with *-h, but this is cohtradicted by all Philippine witnesses, which can indicate only *-q or *-ʔ.

sky:   space between earth and sky

slap, clap

Probably best treated as a convergent development with the common monosyllabic root *-pak₁ ‘slap, clap’.

slice

Probably a chance resemblance.

slide, glide

Based on the above data (except for Old Javanese) Dempwolff (1938) reconstructed *suRsuR ‘to slide, glide’. However, I have not found his Tagalog form in either Panganiban (1966) or English (1986), and even if it is accepted, there is too little semantic agreement among any of the forms cited here to permit a confident reconstruction.

slip, slide

Chance. Dempwolff (1938) proposed Uraustronesisch *linsad ‘to slip, slide’ (Ausgleiten), but I cannot find either the Malay or Javanese forms in any modern dictionary, and the residual comparison is feeble.

slide off:   glide off, slide off

Dempwolff (1938) proposed *sibsib ‘glide off, slide off’, but his comparison is both semantically and phonologically unconvincing. Neither the Old Javanese form (substituted for Dempwolff’s Javanese s<əl>isib, which I cannot find in Pigeaud 1938 or Horne 1974) nor Fijian sisi has any obvious semantic connection with the Polynesian words or with one another, at least under the meaning ‘glide off, slide off’. Moreover, the latter reflect POc *sipo, and are thus morphologically hi-hifo, si-sifo, not hihif-o, sisif-o.

sling, hurl

Late innovation, with some borrowing for Malay, Toba Batak, Dairi-Pakpak Batak, Sundanese, Old Javanese; chance for Gilbertese.

slip, slide

Chance. Dempwolff (1938) proposed Uraustronesisch *linsad ‘to slip, slide’ (Ausgleiten), but I cannot find either the Malay or Javanese forms in any modern dictionary, and the residual comparison is feeble.

slip off

On the basis of this comparison Dempwolff reconstructed ‘Uraustronesisch’ *lurus ‘slip off’. However, the Javanese form appears to contain the adversative passive circumfix ka- -en, found in a number of other forms, making the base milurus rather than lurus. Pigeaud (1938) gives KAMI ‘to pine away’ (wegkwijnen), with no morpheme division, while Dempwolff, citing the earlier work of Jansz (1913) gives the somewhat different sense ‘have a panic attack’ (von Schreck befallen sein). In any case this proposed cognate set shows little semantic coherence, and is best treateed as a product of chance.

slope, mountain side

slow:   weak, slow

slow

Despite its excellent semantic match and near-perfect formal match, this comparison is only possible by accepting a unique correspondence of d- in Tagalog and other Philippine languages, to l- in Malay. This comparison was tentatively accepted in Blust (1970), where it was noted, however (152), that “Ml. l < *d in this form has no known parallels. While chance cannot be completely ruled out as an explanation of the formal and semantic resemblance of these items, the probability seems high that the comparison is valid. An explanation for the discrepant initial segment of the Malay word has not yet been found.” Today, 43 years later, we still lack an explanation.

small

smelt:   melt, smelt

smoke:   rise (of smoke)

smoke:   rise (of smoke)

smoke

sneeze

Convergent innovation. Most of the forms cited here show multiple iconicity, beginning with a low vowel which approximates the position of the mouth at the onset of a sneeze, proceeding through an alveolar or palatal fricative or affricate, which mimics the oral closure preceding the release of a sneeze, and ending with a high vowel (plus nasal), which mimics the narrowed oral aperture and partial nasal channel which typically terminates a sneeze.

sneeze

Probably products of chance convergence motivated by universal sound symbolism.

sneeze

snout, nose

so, so that

so, so that

soaking wet

soil:   dig up the soil

soil:   dig up the soil

song:   sing, song

Chance. Dempwolff (1934-38) compared these forms, but was forced to recognize a fossilized affix in the Malagasy word (h-ira), Tagalog and Bikol híla appear to be connected with híla 'pull, drag, tow, haul; a load, as a unit measure for haulage'.

song

soot

Probably a convergent semantic innovation from PMP *paRa 'storage shelf', since the latter was located above the hearth. In this connection cp. Malay araŋ para ('soot' + 'storage shelf') 'soot'.

sore, scab

sound:   dull resounding sound

Probably convergent innovation from a common root *-beŋ₃ 'dull resounding sound'.

sound:   throaty sound

Tae' iʔduk and Manggarai iruk/ may contain a variant of the root *-dek 'hiccough, sob'. Whether this turns out to be the case or not, the similarity of these forms is best attributed to convergence.

space between earth and sky

space:   gap, space, void

Based on this comparison Dempwolff (1938) posited Uraustronesisch *puhaŋ ‘empty’ (leersein), citing the gloss of Fijian vua as ‘uninhabited’ (unbewohnt sein).

space within a concave object

Based on this comparison Ross (2003:110, 239) proposed POc *loto ‘space within a concave object’. However, the expected Loniu reflex of a POc *loto would be **lot, not lɔtiye, and Polynesian reflexes suggest that there are two distinct cognate sets, one reflexting *loto ‘inner self, feelings, mind’, and the other ‘pool, possible inside the reef’. The former is reconstructed based on evidence from Vanuatu and Polynesia, but any attempt to propose a reconstruction for the latter earlier than Proto-Central Pacific (or perhaps more realistically, Proto-Polynesian) is currently premature.

sparse:   loose sparse roomy

A product of chance. Dempwolff (1938) posited ‘Uraustronesisch’ *luga(r) ‘sparse, not dense’ (undicht).

spicy tasting

Dempwolff used the Toba Batak, Ngaju Dayak and Malagasy forms to reconstruct *siak ‘spicy tasting’. However, the Toba Batak form does not appear to be cognate with those in the Barito languages, leaving us with unrelated reconstructions in Proto-Batak and Proto-Barito. Despite this shortcoming a gloss was assigned based on the meaning in a single witness.

spill out

spit

spite:   provided that, in spite of

split

split down the center

split:   chop, split, adze something

The Thao form could reflect either *saqsaq or *taqtaq, the Paiwan form only *saqsaq. Probably a chance resemblance.

split tear

The Oceanic members of this comparison may be related, but until better evidence is available the resemblance between all of these forms is treated as a product of convergence.

sponge

spread out

spread out

spread out

Mills (1981) used this comparison, together with several other forms from central and southern Sulawesi meaning ‘move, spread, be common, well-known’, to propose PAn *laylay ‘spread, move, go around’, noting that Dempwolff (1938) had assigned the Tagalog form to his *DayDay ‘spread out’, “where it fits very poorly with Javanese ḍeḍe ‘sun oneself’.” However, one must ask whether the comparison proposed by Mills is any more convincing than the one proposed by Dempwolff.

sprout, plant, shoot

Dempwolff (1938) cited Fijian i-so-somi ‘seedling’, but this is not the sense of this term in Bauan Fijian as given by Capell (1968), or in Wayan as given by Pawley and Sayaba (2003). In the absence of further evidence this comparison is best treated as a product of chance.

squirrel

Borrowing from Malay, or a late innovation in Borneo.

stage:   growth stage of rice

stalk

The Fijian and Samoan forms are assumed to be related; the similarity of the Acehnese form to these is due to chance.

stand apart

stand open

On the basis of this comparison Dempwolff (1938) proposed Uraustronesisch *la(m)bas ‘stand open’ (öffenstehen). It seems much better treated as a product of borrowing (Toba Batak from Malay) and chance (the rest).

stand up

Dempwolff (1938) proposed this comparison, which seems to be a product of chance.

steadfast, firm

Dempwolff (1938) reconstructed ‘Uraustronesisch’ *tegeŋ ‘steadfast, firm’ (standhaftsein).

steady:   constant, steady

Dempwolff (1938) proposed *kekel ‘constant, steady’, but the evidence he presented provides little convincing support for his reconstruction.

steep:   aslant, steep, precipitous

Dempwolff (1938) proposed *si(dD)aŋ ‘steep, precipitous’, but this comparison appears to be confined to Ngaju Dayak and Malagasy. 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.

stench

stench

Although this comparison was accepted in Blust (1989) it now appears to be arbitrary, since Nggela aŋo shows parallel sound correspondences and a generic semantic similarity to several forms in non-OC languages. For a general discussion of the word family *qaŋ- meaning 'stench' cf. Blust (1988:60ff).

stench of fish or blood

The Ngaju Dayak, Malay and Javanese forms are cognate, but probably represent a late innovation in western Indonesia. The resemblance of the Tagalog and Malagasy forms to these is assumed to be a product of chance, and although the South Sulawesi forms appear to be cognate, they point to an etymon with penultimate *e, and so do not permit a clear-cut reconstruction.

step

step onto

still, calm

This probably is best treated as a collected of unrelated forms until such time as better comparative data is available. Dempwolff (1938) reconstructed Uraustronessisch *liŋaw ‘still, calm’ (Stillsein).

still, yet

Probably a product of chance. Dempwolff (1934-1938) reconstructed ‘Uraustronesisch’ *maniŋ ‘still, yet (additive)’

stink

stink

Chance. Ngadha beŋu probably is connected with Manggarai beŋur 'odor of something rotten', beŋus 'rotten (of smell)' or beŋut 'somewhat rotten (of a gourd, etc.).

stink

stir up a fire

Dempwolff (1938) cites this comparison, but I find nothing like the Tagalog form in any modern dictionary of the language (it is more likely to be káhig ‘scratch up the soil, as a fowl’). Zorc (1971) cites this for PPh with the meaning ‘fire’, but so far as I have been able to determine it is entirely without support.

stout, stocky

stomach

The New Ireland forms are related; their resemblance to the similar items in Taiwan

stop, cease

stop, of the wind

stop for

Chance. For the source of Malagasy áfitra cf. *ampet.

stop by, to moor a boat

Based on this comparison Dempwolff (1938) proposed ‘Urindonesisch’ *tuli ‘to moor; stop by to visit’ (Einkehren, Anlegen).

stout, stocky

stratum:   layer, stratum

stream, pour

stream of water

On the basis of the Malay and Malagasy forms Dempwolff (1938) proposed *panca(rR) ‘stream of water’ (Wasserstrahl), but until a better comparison is available this one, even with the addition of Isneg pásag, is best treated as a product of convergence.

stretch:   obstruct, stretch across

stretch out

Dempwolff (1938) reconstructed *kuju(r) ‘stretch out’, but there is little to distinguish this from a chance resemblance.

strew around:   disseminate, strew around

Dempwolff (1938) reconstructed *seba(r) ‘disseminate, strew around’, but plausible cognates are found only in western Indonesia, where they may be borrowings from Malay. Fijian cova does not appear in Capell (1968).

strike:   collide, hit, strike

The Philippine forms are unrelated to those in western Indonesia. Of the latter only Bidayuh (Bukar-Sadong) and Malay, or Malay and the Batak languages permit a comparison, and this cannot safely be attributed to PWMP.

string:   to string fish, etc.

Chance. Gilbertese itu may be cognate with Motu Jituri 'a string of fish' (< POc *i-tuRi 'instrument for stringing'), but this interpretation cannot be applied to Nggela itu-itu.

stripe

stroke (as in swimming)

stroke scratch

Dempwolff posited *kaDus ‘stroke, scratch’ based on these two forms, which appear to have no connection with one another.

strong tasting

Proposed by Dempwolff (1938) as ‘sharp tasting’. The Ngaju Dayak form probably is borrowed from Banjarese, and the similarity of these to the Tagalog word is best attributed to chance.

stupid

stupid

substance used in fermenting palm wine

Apart from Toba Batak raru, which probably is a Malay loanword, the other forms cited here appear to show a similarity to one another that is due to chance. Dempwolfff (1938) proposed Uraustronesisch *laru ‘substance used to purify fermenting liquids’ (Mittel zum Klären gärender Flüssigkeiten).

suck

suck

Probably a convergent development.

suffix:   numeral suffix

sun:   day, sun, light

Chance. Kähler (1961) cites this comparison, but Mongondow y can only reflect *d, *r, or *y, none of which are regular sources of Simalur, Nias -l-.

support from below

supposition

suspect:   guess, suspect

suspend from poles

swallow:   to swallow

sway

Makassarese boyoŋ appears to be a loan from Malay. The resemblance of the other forms both to this word and to one another is regarded as a product of chance, although the sequence -VyV- is known to be symbolic of swinging or swaying movement in other morphemes (Blust 1988:57-58).

sway

Based on this set of data Dempwolff posited *bayaŋ ‘sway’, but the resemblance between these forms is best attributed to chance. As noted in Blust (1988:57-58), medial –y- is common in morphemes meaning ‘swing, sway, vacillate’ and the like.

sway, move

sweep away

swimming:   stroke (as in swimming)

swing, wave

swollen