The PAN coda *-X in Favorlang

In Sagart (2019 here) I reconstructed a new PAN consonantal ending : *X [χ]. This is one source of the Kra-Dai tone B. In the same paper I showed that at the end of Austronesian words, two previously reconstructed h-type sounds: *H1 and *H2, evolved to different tones in Kra-Dai: *H1 to tone C and *H2 to tone A. In Table 7 of that paper I further assigned my *X to two words having no Kra-Dai cognates: *qeCaX ‘chaff’ and *duRiX ‘thorn’.  The reason for claiming these words ended in *X was that, while there was evidence in each case for a final voiceless fricative (Amis ʔtah ‘chaff’, Aklanon dugih ‘thorn’), that fricative could neither be *H1 nor *H2, as shown by the combination of  cited Saisiyat, Amis and Bunun forms: only *X was consistent with all the evidence.

Favorlang, an extinct Austronesian language of the Formosan west coast, recorded by Dutch missionaries in the mid-17th century brings additional evidence. I rely on Happart’s Favorlang vocabulary of 1650, as published in Campbell (1896).

As represented in those sources, Favorlang regularly loses the PAN endings *H1 and *H2:

*-H1: bao ‘new’ < baqeRuH1; bagg’ (o chau) ‘a live coal’ < baRaH1; ema < *qumaH1 ‘a field’, atshia ‘a salting’ < qasiRaH1.

*-H2: torro ‘three’ (in manna-torro-us ‘three times’) < tuluH2; manna-pida ‘to be often repeated’ < *pijaH2; charina ‘ear’ < CangilaH2; abo < qabuH2 ‘ash’; kairi ‘left’ < ka-wiRiH2; bato ‘stone’ < batuH2.

Yet *qeCaX ‘chaff’ is reflected in Favorlang with a final back-of-the mouth fricative, as chach or chagh ‘the powder of ground seed; the refuse of bruised corn’ (understand ‘the by-product of pounding grain: chaff’). PAN *q is regularly lost in Favorlang, and PAN *C evolves regularly to ch. Dutch orthographic ch writes [x] and gh writes [x] or [ɣ] : this may have been close to the value of the fricatives in chach/chagh: [xax], [xaɣ] or similar. The first ch- reflects PAN *C; the last consonant: –ch or –gh, my *X. It cannot reflect *H1 or *H2, since as shown above both go to zero in Favorlang.

The behaviour of final *-X in Favorlang is parallel to that of final -h in the Amis cognate ‘tah ‘chaff’: this -h cannot reflect *H2, as final *H2 goes to zero in Amis. It could reflect *H1, since Amis reflects final *H1 as –h, but *H1 is contra-indicated by Saisiyat kæ-ʔsæʔ ‘chaff’—expect ʔsæh if PAN was *qeCaH1. Thus geographically distant Amis and Favorlang agree in having in ‘chaff’ a final back-of-the-mouth fricative that can neither reflect *H1 nor *H2.


Happart, Gilbertus (1650 [1896]) Woordboek der Favorlangsche Taal. Published in English translation as Happart’s Favorlang vocabulary in W. Campbell (ed.) The articles of Christian Instruction in Favorlang-Formosan Dutch and English. London, Kegan Paul, Trench and Trübner, 1896, pp. 122-199.

Sagart, Laurent. 2019. A model of the origin of Kra-Dai tones. Cahiers de Linguistique Asie Orientale, 2019, 48 (1), pp.1-29. ⟨10.1163/19606028-04801004⟩

Cite this post as: Laurent Sagart, "The PAN coda *-X in Favorlang," in Sino-Tibetan-Austronesian, 24/06/2020,


(view clickable tree)

(jump up one level to Pituish)

A monophyletic Favorlang-Taokas subgroup is supported by these innovations:

1. an innovative numeral for ‘6’: Favorlang and Babuza nataap, Taokas tahap, tákap , takkap (sets #26 and 28 in the ABVD).   These are additive forms based on a 1+5 prototype (*sa-RaCep), different from Saisiyat and Pazeh where ‘6’ is 5+1.

The Favorlang-Babuza form nataap  is composed of natta ‘1’ and achab [ahab] < *RaCep  ‘five’ reduced to [ab]. natta in turn reflects PAN *sa ‘one’ with na- prefixed marker of a derived series of numerals of uncertain functions: na-rroa ‘two’, na-torro-a ‘three’, na-spat ‘four’, na-chab ‘five’, na-taap ‘six’.

Taokas tahap, tákap , takkap ‘six’ is composed of ta- < *sa ‘one’ plus hap, kap, kkap in different dialect points. Initial h, k, kk reflects PAN *R. PAN *C should give Taokas s: here it is lost, presumably after *RaCep was reduced to Rasp through syncope.

2. An innovative word for ‘dog’. Favorlang mado, Babuza mato/malu/mado/malok (dialects), Taokas mato/maro/mazok/marox/malok etc. (dialects), all ‘dog’ (set #9 in ABVD). These forms go back to *ma[d]uq and displace PAN *asu, u-asu ‘dog’.

A Favorlang-Taokas subgroup is common to my phylogeny and to Blust’s.


ABVD (Austronesian Basic Vocabulary Database). Online at Accessed June 19, 2020.

Cite this post as: Laurent Sagart, "Favorlang-Taokas," in Sino-Tibetan-Austronesian, 20/06/2020,

Pituish (master post)

(view clickable tree)

(jump down one level to Limaish)

Pituish has two branches: Favorlang-Taokas (here) and Limaish (here). It is defined by the innovation of *pitu ‘seven’, as  discussed in Sagart (2004) (here), to which readers are referred. *pitu replaces PAN *RaCep-i-tuSa ‘five-plus-two: seven’, of which it is a left- and right-pruned form: *pitu < *(RaCe)p-i-tu(Sa).

In addition, a non-displacing innovation in Pituish created a new numeral for ‘nine’, based on *sa-ŋ-aCu ‘one taken away’ (from ‘ten’):  Favorlang tannacho, Babuza #131 tanahu,Taokas tanasu, Thao: tanacu. It is listed as cognate set #6 in the ABVD (here). All through Limaish and Enemish, the new numeral competed with the inherited *RaCep-i-Sepat ‘five plus four’ found in Pazeh xaseb-i-supat. It did not displace it either in Limaish or in Enemish, since later *Siwa, simplified from *RaCep-i-Sepat, finally displaces all its competitors in Walu-Siwaish. *sa-ŋ-aCu ‘9’ does not appear in  the other Pituish languages upstream of Proto-Walu-Siwaish, because of a series of local innovations displacing *sa-ŋ-aCu:  Atayal qeru, Sediq maŋali, and in Siraya matauda.


Sagart, L. (2004) The higher phylogeny of Austronesian and the position of Tai-Kadai. Oceanic Linguistics 43,2: 411-444.

Cite this post as: Laurent Sagart, "Pituish (master post)," in Sino-Tibetan-Austronesian, 20/06/2020,

‘Bicycle’ to ‘bike’, and the limits of regular sound change.

The development of bike [bajk] out of  bicycle [‘bajsɪkl] has occasioned some etymological speculation in print and online.  The two principal proposals seem to be (a) that bicycle was reduced from a trisyllable to a monosyllable through cumulative loss of its medial syllable [sɪ] and of final syllabic [l], thus bicycle > bicycle  > bic [bajk] (here); and (b) that bicycle was pruned to bicycle , the result being [bajk] rather than [bajs] because the phoneme underlying [s] was a velar /k/ (Hausman 1976).  This last, generatively-inspired proposal is not hugely plausible. If the underlying phoneme was /k/, why was the velar brought to phonetic light after pruning, when the etymological connections of –cycle were least apparent ? why does /k/ never surface in cycle and its derivatives ?

Here is  a third account, which I believe makes better historical sense. It agrees with the first as regards pruning  of final syllabic [l], but differs from it in supposing syncope of the second, unstressed vowel [ɪ], rather than loss of the entire second syllable [sɪ] in bicycle. The two changes, l-pruning and ɪ-syncope,  applying in whatever temporal order, resulted in [bajsk], with [k] resyllabified as part of the monosyllable’s coda. The modern form bike in turn results from the simplification to [jk] of the  cluster [jsk], unattested in word-final position, bringing it in line with like, hike, pike, Mike etc.

There are interesting lessons in this. When polysyllabic words see a marked and sudden increase in their frequency, as for instance following fast societal acceptance of a technological innovation, as here, or for whatever other reason, there will be pressure on them to  become shortened. Shortened forms can take hold very fast. Thus in French, the five-syllable masculine noun coronavirus [kɔʁɔnaviʁus], in use at very low frequencies since c. 1965, suddenly underwent a sharp frequency increase at the beginning of the Covid-19 pandemic in December 2019; I first heard it shortened to trisyllabic [kɔʁɔna] on March 15, 2020 (the new trisyllable was itself subjected to competition from disyllabic [kɔvid] shortly afterwards, in April or May, 2020, and at the time of writing, [kɔvid]  appears to have gained the upper hand). Meanwhile in the UK, a parallel reduction of coronavirus to disyllabic rona appears to have been underway, as shown by this  journal article dated July 28, 2020.

The main tools language has at its disposal to shorten words are pruning and unstressed vowel syncope. One works from a word’s edge—or both edges, as in rona—, the other from the inside. Both affect words that are too long for their frequency,  without any stateable phonological conditioning.  Syncope will result in unattested or unacceptable consonant clusters, and prunings may affect syllabification, in ways that require repairs in order for the shortened word to fit the language’s canon and phonotactics. Strings will be resyllabified, clusters simplified. As the clusters will be rare or unattested, they will tend to simplify in unique ways. Although they cannot be regarded as ‘regular’, these are common historical changes affecting the sound of words.

  • Hausmann, R. B. (1976). An Etymological Brainteaser: The Shortening of Bicycle to Bike. American Speech, 51(3/4), 272. doi:10.2307/454976 
Cite this post as: Laurent Sagart, "‘Bicycle’ to ‘bike’, and the limits of regular sound change.," in Sino-Tibetan-Austronesian, 19/06/2020,