*k   

*kabaR clothing

*kabus₂ shrimp, lobster

Note:   Possibly a loan distribution, although this seems unlikely since both the Kavalan and Puyuma are in contact with the sea, and there is no obvious reason why the word for such a familiar creature would be borrowed. However, PAn *qudaŋ must be reconstructed with the same meaning based on cognates in Paiwan and many Malayo-Polynesian languages, raising questions about why two terms would coexist with the same reference.

*kahuS scoop up

Note:   Also Pazeh mu-haus ‘to scoop up, as noodles’. Li (1978) gives only the Actor Voice form for Saisiyat, and the base cited here is extracted without the benefit of direct attestation. This word shows -s for expected -ʃ, but seems clearly to be related, and provides evidence for *h, which in general can be obtained only from reflexes in Saisiyat, Pazeh, or Amis.

*kakaCu spider

Note:   Also Thao, Tsou kakatu ‘spider’, both of which appear to be loanwords from Bunun.

*kaluŋkuŋ deep resounding sound

Note:   With root *-kuŋ₂ ‘deep resounding sound’. Possibly a convergent innovation from the same monosyllabic root.

*kapua silk cotton, kapok

Note:   The glosses for both of these languages state that the material in question is ‘cotton’, but it is clear from the ethnobotany of Taiwan that *kapua refers to the fibers of the silk-cotton tree, Ceiba pentandra, a native species on the island. Although Chen (1988) does not touch on it, other sources state that kapok was traditionally used as a lining in woven materials to keep out the cold.

*kawaS₁ year, season; sky, heaven

Note:   PAn *kawaS is in competition with *CawiN ‘year’ which, however, is reflected only in southern Taiwan. It is difficult to achieve an adequate gloss for this word, as it apparently represented a concept for which there is no English equivalent, one which combined the notion ‘sky’ as the abode of deities rather than a physical domain (cf. *laŋiC ‘sky’), and at the same time related this feature of Nature to the cycling of the seasons. This word was replaced in PMP by *taqun.

*kawaS₂ talk, talk about, discuss

*kaway₃ net bag used by men for carrying things on back

Note:   Also Kavalan qaway ‘to carry a baby in a sling/cloth on the side’.

*kawkaw₁ sickle for harvesting grain

Note:   There is some uncertainty about the PAn status of this form, given the semantic latitude that is allowed in admitting Thao kawkaw into the comparison.

*keRiw hemp plant; hemp

Note:   Up to the mid-twentieth century some Formosan aborigines still used hemp to weave clothing. To prepare it for weaving it was first soaked in water to soften the fibers.

*ketun cut, sever

*kiRim seek, search, look for

Note:   This comparison was first described by Tsuchida (1976:290).

*kuCkuC roll up, curl up tightly

*kulay₁ worm, insect

Note:   Also Thao ku-kulay ‘generic for creepy-crawly creatures, including beetles, worms, and any unidentified insect or other small terrestrial invertebrate; bug’. The Thao word is clearly borrowed, presumably from Bunun, although lexicographic resources for Bunun are inadequate to determine this with certainty. It is possible that this was a PTSo innovation that spread into both Siraya and Thao.

*kumeS pubic hair

*kumiS pubic hair

Note:   Also Saisiyat romiʃ ‘beard, moustache’, Kavalan umis pubic hair. The meaning of the PAn form is based on the semantic agreement between Thao and the preceding forms, which show an irregular reflex of the initial consonant in Kavalan, and of the second vowel in Amis. The meaning of this form apparently had shifted to beard, moustache in PMP, although a PMP reconstruction for pubic hair is yet to be proposed.

*kuNkuN Formosan hill partridge

*kuRi rake

*kuti vulva, vagina

*kuya defect, bad, defective

Note:   I have cited an exceptionally large number of affixed forms of this base in Paiwan in order to better tease out the semantic range of the concept designation by the term. Given its very limited attestation it is possible that this is a chance resemblance, but the evidence we have suggests that the comparison is valid. This form was replaced in PMP by *zaqet. This comparison was first proposed by Tsuchida (1976:175).