*b   

*baCaR broomcorn millet: Panicum miliaceum

Note:   Also Malay batari ‘Andropogon sorghum’. The following more restricted reconstructions have been proposed: Proto-Central Philippines *batad ‘a plant: Andropogon sorghum’ (Zorc 1979:43), "Proto-Indonesian *bataD ‘grain sp., millet’ (Mills 1981:65). In addition, Mills notes the material earlier assembled by Heyne (1950). The correct interpretation of the shape and meaning of this comparison is due to Laurent Sagart (p.c.).

*bajiw edible mushroom sp.

Note:   Based on data from Puyuma, Kavalan and Amis, Li (1994) posited ‘Proto-Southern Formosan *baɬiw’ (= *baNiw). However, Paiwan vadiw cannot be reconciled with this reconstruction.

*baki₂ grandfather

*bali₃ wind

Note:   This was clearly the PAn word for ‘wind’, and was replaced in PMP by *haŋin, a form that is widespread throughout the Austronesian language family outside Taiwan.

*baliw₃ buy, sell

*baŋaS a tree: Melia azedarach

Note:   This comparison was first proposed by Tsuchida (1976:235). Li (1994) posits ‘Proto-Formosan’ *baŋas ‘Melia azedarach’ for what he evidently intended as *baŋaʃ.

*baŋeS skin

Note:   Although it is sometimes assumed, following Dempwolff (1938), that PAn *kuliC meant ‘skin’ (e.g. Tsuchida 1976:225-26, Zorc 1995:1185), this word apparently referred to the peelings of fruits and tubers, and only shifted semantically to ‘skin, bark’ in PMP. The present form, which has previously been completely overlooked even though the Saisiyat and Amis forms appear on the same page in Ferrell (1969:237), appears to be a better PAn candidate for the meaning ‘skin’, since borrowing offers a poor alternative to common inheritance. Like a number of other PAn reconstructions, both those with reflexes confined to Formosan languages and those with reflexes in Formosan and MP languages, this comparison attests the high level of linguistic diversity within Taiwan, which has led to some inferrable PAn forms being retained in only one or two Formosan languages.

*baŋun the Japanese cypress: Chamaecyparis obtusa

*baqeSiŋ sneeze

*baRaŋ rib (?)

Note:   Also Thao falhán ‘rib’, with irregular final stress. The meaning of *baRaŋ is unclear. While it evidently referred to a part of the body that included the region of the thoracic cavity, it is in competition with the better-supported PAn *tageRaŋ in the meaning ‘ribcage’. This comparison was first brought to my attention by Laurent Sagart.

*baRat₂ cucumber: Cucumis sativus

Note:   Reflexes of *baRat appear to be restricted to these two languages, and the term was replaced in PMP by *qatimun/katimun. This comparison was first noted by Li (1994).

*benan deer of the genus Cervus, either the sika deer or sambar deer

Note:   The referent of PAn *benan was clearly distinct from the muntjac or barking deer (PAn *sakeC), but the two witnesses for which we have explicit glosses differ in which of the two larger Cervidae they indicate. This term thus served either as the name for both the sika and sambar deer, or as one of the two that cannot be determined from the available evidence.

*beNbeN banana: Musa sapientum L.

Note:   Based on Basai, Trobiawan puti ‘banana’ and cognates in Malayo-Polynesian languages reaching as far east as Fiji Wolff (2010) suggests that only PAn *puti can be reconstructed in the meaning ‘banana’. This inference runs against the observation that reflexes of *beNbeN are found in at least some members of every primary branch of Austronesian that is represented in Taiwan (Atayalic, Northwest Formosan, Western Plains, Bunun, Tsouic, Rukai, Puyuma, Paiwan, East Formosan), making this a far stronger candidate.

It must be kept in mind that the Spanish were present in the Taipei basis from 1626-1642, and left loanwords such as baka ‘cow’ (< Span. vaca) and paskua ‘New Year’s Day’ (< Span. Pascua ‘Easter’) in Kavalan. Since the Spanish came to Taiwan from Manila and almost certainly brought Filipinos with them, the typical Malayo-Polynesian words for ‘banana’ in Basai and Trobiawan probably are best attributed to borrowing of a form like Pangasinan punti ‘banana’. Alternatively (since reflexes of PMP *punti are rare in the Philippines, and absent from Tagalog), it is conceivable that both *beNbeN and *puti were found in PAn, the first as a generic term for Musa spp. and the second as the designation for a particular banana species or variety that expanded its field of reference and acquired a preconsonantal medial nasal in PMP. As with other reduplicated monosyllables, it is assumed that the medial consonant cluster in this word was broken up by schwa epenthesis in the separate histories of several Formosan languages, a hypothesis that is further supported by the occurrence of a word-final schwa in Rukai and the Tsouic languages Kanakanabu and Saaroa, which do not permit unambiguous final consonants.

*beRuS a plant: Rhus semialata

Note:   Part of this comparison was first noted by Li (1994) who, however, was unable to achieve an adequate reconstruction.

*biliN to follow after someone, walk behind

Note:   Tsuchida (1976:141) proposed ‘Proto-Hesperonesian’ *biliN, comparing these Formosan forms with Philippine forms such as Itbayaten vilin ‘request, order, command, message, instruction’, ma-milin ‘leave behind’, Hanunóo bílin ‘order, request, requisition, command’, Hiligaynon bílin ‘request, last words’, mag-bílin ‘to enjoin, to leave behind’, and Cebuano bílin ‘leave something behind; for a wife to be left pregnant when her husband dies’; one’s turn to be left behind; words, orders left by someone; someone who stays behind’, bílin ug púluŋ ‘leave word , pa-bílin ‘stay behind; remain in the same way’. These forms may all belong to a single cognate set, but until clearer evidence of a connection between them emerges they are treated here as distinct.

*bukuN back (anat.)

Note:   This form is in competition with the better-attested PAn *likud for the meaning ‘back (anatomical)’. Whatever semantic nuance may have distinguished these terms remains unclear.

*bula₂ to give

Note:   Although this comparsion appears to be both phonologically and semantically straightforward it is in competition with the far better-supported *beRay, and so may be a product of convergent innovation.

*bunuR₁ to swell; swollen

*bunuS machete

Note:   The antiquity of this form is uncertain, as it appears to imply the use of metal to forge tools. To date it is known only in Thao and Amis, which is both an asset and a liability for the reconstruction. The liability arises from its relatively weak attestation, but the asset arises from the unlikelihood of borrowing between languages that are not in contact, or have been in contact within the recorded past.

*buRay flower

Note:   Also Atayal buay ‘fruit’, Kavalan muray ‘flower’. This comparison was first proposed by Dyen (1995:486), who commented that the loss of *R in Atayal “is one of a small number of such instances.” If valid, I assume that *buRay meant ‘flower’, since PAn *buaq ‘fruit’ is far better supported.