*i   

*-i₁ imperative suffix

Note:   Egerod (1965) describes three "passives" for Atayal: (1) a "relational passive", (2) an "indefinite passive", and (3) a "definite passive". The definite passive can occur in any of five aspects, of which the "neutral definite passive" is marked by -an, while the "negatable neutral definite passive" and the "imperative definite passive" are marked by -i. The paradigmatic alternation of -an and -i thus appears to be common to at least Atayal and Bikol. The present affix evidently is identical to what Wolff (1973:73) called the "local passive dependent" construction. It seems to have marked imperatives of a certain type, but may have been used with a wider range of construction types than I have indicated here.

*-i₂ local transitive suffix

Note:   Pawley (1973) initiated the discussion of this affix, which he called a ‘close transitive’ suffix, by drawing attention to similar functions which it performs both in Oceanic and in non-Oceanic languages. Starosta, Pawley, and Reid (1982:155) suggest that it derives from the generic locative marker *i which under certain conditions lost its prepositional function and was incorporated into the preceding verb in the separate histories of many Austronesian languages. I concur with this interpretation, but note that it remains unexplained why the incorporation of such a preposition would not occur with intransitive as well as with transitive verbs.

Judging from its widespread use as a verbal affix in Oceanic languages, it appears that the preposition *i had already become a suffix in POc (though at the same time it continued to function as a generic locative marker). Because reflexes of *-i appear to be less common in non-Oceanic languages, and because they sometimes co-occur with an active verb prefix (as in Malay and Toba Batak), it seems likely the verbal functions of *-i in Wolio developed independently from those in Malay and Toba Batak -- in other words, that a suffix *-i ‘local transitive’, as distinct from the preposition *i, did not exist in PAn or PMP. Moreover, in Oceanic languages the local transitive suffix *-i so commonly co-occurs with the incorporated 3sg. object pronoun *-a that the entire sequence *-ia (q.v.) must be posited as a functional unit for POc. Starosta, Pawley, and Reid appear to conflate *-i ‘local transitive’ and PAn *-i ‘imperative suffix’, which I treat as separate.

*i₁ exclamation of wonder, disgust, etc.

Note:   Also Bunun ia ‘exclamation’, Lakalai hi ‘exclamation of excitement’ points to POc *qi, but is contradicted by the lack of an initial consonant in Paiwan i. The creation and shaping of the above forms clearly has been subject to the operation of universal human tendencies, but this by itself is no reason to assume that the forms compared have been independently invented in all daughter languages.

*i₂ generic marker of location in space or time

Note:   I take the glottal stop in Tongan ʔi to be secondary, as it disagrees with the cognate forms both in Rennellese and in Paiwan (for further discussion of the introduction of historically secondary glottal stop in Tongan grammatical formatives cf. Clark 1976 and Harrison 1991:135, fn. 5). PAn *i evidently was a generic marker of location which has become fossilized in most reflexes of *i-ni ‘this, here’, *i-tu ‘that, there (2p.)’, and *i-na ‘that, there (3p.)’, and in many reflexes of *i babaw ‘above’ and *i babaq ‘below’, but which continues to be used productively in many languages before other nouns. For further details cf. Blust (1989).

*i₃ personal article

Note:   Further documentation of *i ‘personal article’ is found in Blust (1977), where it is argued that *i and *si were in complementary distribution, at least when attached to pronouns (*si marking 3p., *i marking non-3p.). In many attested languages the distinction between this morpheme and *qi ‘genitive marker’ is obscure.

*i₄ numeral prefix

Note:   Also Bauro (Tryon and Hackman 1983) irua ‘two’ (the only Bauro numeral recorded with i-).

*-ia transitive and imperative suffix

Note:   Pawley (1973:120ff) has suggested that this form contains two morphemes: 1. *-i ‘close transitive’ and *-a ‘3sg. object’. As partial justification for this analysis he points out that reflexes of *-a are suffixed directly to the verb in Bauan Fijian and, under certain conditions, in other Oceanic languages such as Motu and Roviana. While this analysis may be correct, the combined form *-ia is so widely reflected in such similar functions that it can be reconstructed independently of *-i or *-a.

Two facts associated with this etymon appear especially noteworthy:

(1) Apart from the well-known problems associated with the "thematic" consonant that surfaces before reflexes of *-ia in many languages, Motu and Roviana agree in showing surface -ia only after stems that end in a; following stems that end in e, i, o, or u, the semantically similar surface form is -a (Motu: ita-ia ‘to look at’, but lare-a ‘to shoot an arrow’, isi-a ‘to husk a coconut’, hao-a ‘to awaken, arouse’, utu-a ‘to draw water’; Roviana: asa-ia ‘to grind, as an axe’, lage-a ‘to fix a parrot to a perch’, gili-a ‘to tickle gently’, tapo-a ‘to grope or feel for, as in the dark’, vatua- ‘to give’). I know of nothing which would produce such an agreement through convergence, and therefore assume that in POc there was an alternation between *-ia, following stem-final *a, and *-a, following other stem-final vowels. One or the other variant appears to have been generalized in various languages; for convenience the two will be referred to as *-ia.

(2) Reflexes of *-ia function both as transitive suffixes and as imperative markers in several widely separated Oceanic languages. In Eastern Polynesian languages reflexes of *-ia typically mark passive constructions, and also appear in certain types of imperatives (those that imply a goal or object). This association of passive and imperative appears to closely parallel the association of passive and ‘polite’ imperative associated with reflexes of *-en ‘goal focus, direct passive focus’ in many WMP languages. While no one has yet argued that POc had a passive construction marked by *-ia, it apparently had an imperative construction marked in this manner. The fact that the dual function of passive and imperative is associated with reflexes of PAn *-en in some WMP languages and with reflexes of POc *-ia in some OC languages raises the question whether the passive function of *-ia might be far older than PEP.

*ia₁ 3sg. personal pronoun: he, she; him, her, it

Note:   Also Malay dia ‘3sg.’, Uma hiʔa ‘3sg.’, Bare'e siʔa ‘3sg. and pl.’. For arguments supporting *si ia (rather than *s-ia), cf. Blust (1977). Although no reflexes of *ia are known in Formosan languages, reflexes of *si ia in Atayal, Saaroa, and Bunun clearly imply that *ia was present in PAn, and it is posited here on this basis.

*ia₂ demonstrative pronoun and adverb: this, here; that, there

Note:   Also Sasak iaʔ ‘this’, Rembong iaʔ ‘over there’, Sika iʔa ‘this’, Rennellese ʔia ‘here! here it is! take it! now!’. Like *ia ‘exclamation’, this item almost certainly is identical with the 3sg. personal pronoun. Historically pronouns often develop from demonstratives, but in the present case there can be little doubt that the primary sense of *ia was ‘3sg. personal pronoun’, and that both the exclamatory and the demonstrative meanings are semantic extensions (cf. the last three items above, where the demonstrative meaning is found in a word containing the fossilized personal article *si).

*iag cry out, shout

Note:   Also Proto-Chamic *hia ‘to cry’.

*iak cry out, shout

Note:   Also Bontok iék ‘to laugh, be laughing’.

*ian₁ dehortative: don't!

Note:   Cense (1979) gives Makassarese ia ‘may not’, but notes that this item occurs only as a bound form in constructions with a suffixed pronoun, such as iaŋku, iannu, and ianna/. His sole justification for positing a morpheme boundary before, rather than after the first nasal, appears to be that some vowel-final stems unpredictably show an "accretive nasal" before the pronominal suffix (Mills 1975:1:78, 221). However, such vowel-final stems and stems that end in ŋ appear to exhibit identical suffixed forms. Phonemically, then, Cense's /ia/ could as well be /iang/.

*ian₂ dwell, reside, live in a place

Note:   Also Kadazan Dusun izon ‘to live in a place’, izon-on ‘domicile, living place, place of residence’, Lamaholot iaʔ ‘live, reside, dwell’.

*ian₃ that, there (probably 2p.)

Note:   Also PC *hiã ‘there’. Dempwolff (1934-38) attributed Tagalog ián ‘that’, Malagasy izani ‘this’, Sa'a ie ‘this, these’ to *ian ‘that (one), demonstrative pronoun’. Although he did not note it, the Malagasy form is irregular, and can better be compared with Atayal iani ‘this; here; now’, Mansaka y-ani ‘this’, Nggela iani ‘here’, probably reflecting *ia-ni ‘this; here’. The inclusion of Sa'a ie in Dempwolff's comparison is equally problematic. Many OC languages which have lost original final consonants (e.g. Tolai, Bugotu, Mota, Tongan, Niue, Samoan, Anuta, and Maori) have a demonstrative pronoun reflecting earlier *ia. While some of these forms can plausibly be compared with reflexes of *ian, they can equally well be compared with e.g. Aklanon iá ‘here’ or Rotinese ia ‘this; here’, which cannot reflect a consonant-final etymon. As suggested by some grammarians and lexicographers (e.g. Churchward 1959) these demonstrative uses, like the widespread exclamatory and confirmatory forms of similar shape, probably represent semantic extensions of the 3sg personal pronoun *ia.

*iap kind of brown fish

*iaqi demonstrative marker of uncertain meaning

Note:   Also Rotuman ia ʔi ‘this person here’ (cp. ia ʔo ‘that person there’). Ross (1988:208ff) reconstructs POc *iai as a "locative proform", widely reflected in the Papuan Tip languages of southeastern New Guinea.

*ias₁ particle of doubt or interrogation

Note:   Possibly a product of convergence.

*ias₂ maggot, fly egg

Note:   It is unclear how the meaning of this term differed from that of *qulej. It is possible that *ias referred to the eggs of flies prior to hatching, whereas *qulej clearly referred to maggots, caterpillars, and other insect larvae that metamorphose into adults that fly.

*iba another, other; different, distinct

Note:   Possibly identical to PAn *ibaS ‘companion; close relative; another one’.

*iba iba remainder, surplus

Note:   Also Bontok ibʔá ‘remainder, rest’. Possibly connected with PAn *ibaS ‘companion’.

*iba₁ self

*iba₂ bad, evil

Note:   Possibly also Iban antu iba ‘demon, spirit (of the forest?)’. If the latter is included in this comparison the etymon might include an additional gloss ‘bewitched’.

*ibak break or split off

Note:   With root *-bak₃ ‘split off, separate from’.

*ibaS companion, close relative, other one

Note:   Also Bontok ibʔá ‘other; companions’, ibʔa-an ‘join with, put together with’, Kankanaey ibʔá ‘companion, fellow, associate’, Tiruray iboʔ ‘sister-in-law (m.s.), brother-in-law (w.s.)’. Based on the comparison Tagalog ibá ‘other, another, different’, Toba Batak iba ‘person, self’, leban (< *la-iba-an) ‘other; excess, remainder, stranger’, Javanese ébah ‘to move’, éwah ‘to change’, Ngaju Dayak iwa-n ‘(the other =) sister-in-law (w.s.)’ Dempwolff (1934-38) reconstructed *ibaq ‘different; to change’. The present comparison includes the Tagalog and Ngaju Dayak members of Dempwolff's comparison, but not the Toba Batak and Javanese members.

Although I believe that it is an improvement on Dempwolff, the cognate set proposed here is difficult for two reasons: (1) in most reflexes the proposed suffix *-an is fossilized (whereas the prefix *ka- in *ka-ibah-an is fossilized only in Penan k-ivan); (2) the meaning of reflexes ranges over ‘companion, friend’, same generation consanguines, same and adjacent generation affines, ‘clan’, and (at least in Philippine languages) ‘other, different’. I assume that all of these senses are related through the notion ‘companion: the "other" in a personal dyad which includes the self’. Dempwolff presumably would also have included forms that I assign to *iba iba ‘remainder, surplus’. It is possible, but by no means certain,

*ibe mat

*ibeR saliva in the mouth; drool; desire, crave, lust for

Note:   Also Kapampangan ibug ‘have a taste for, be eager to’. PMP *ibeR evidently referred both to saliva in the mouth and to strong desire associated with "watering at the mouth". The meanings of most of the affixed forms are uncertain.

*ibit touch, grasp

Note:   With root *-bit₁ ‘hook, clasp; grasp with fingers’.

*ibo edible sea worm: Sipunculus spp.

Note:   Gilbertese ibo may be a Polynesian loan, but the comparison can stand even if this form is eliminated. This item appears to be distinct from PMP *ibaw ‘edible shellfish’.

*ibu₃ container for liquids

Note:   Gilbertese ibu ‘calabash, gourd, toddy shell made of coconut shell’ is assumed to be a Polynesian loan.

*ibuk dull, muffled sound

Note:   The Iban and Karo Batak forms permit a PWMP *ibuk ‘bird with deep coo or hoot’. Apparently with root *-buk₂ ‘pound, thud, heavy splash’.

*ibun bird sp.

Note:   Possibly a Tagalog loan in Ilokano, but the more specific gloss in the latter language, and the indication that it is obsolescent argue against this interpretation.

*ibus₁ the gebang palm: Corypha gebanga

Note:   Also Kapampangan ebus ‘the gebang or cabbage palm: Corypha utan Lam. Arecaceae (Madulid 2001).

*ibus₂ dessert made of rice

*ibut₁ pull out, uproot

Note:   With root *-buC ‘to weed, pluck, pull out’.

*ibut₂ breeze, draft of wind

Note:   Possibly a convergent resemblance. In Kadazan Dusun *-b- normally became v; alternatively, Kadazan Dusun ibut may be a reduced form of earlier *imbut.

*icak step, tread on

Note:   Also Balinese iŋsak ‘trample on, rub the feet on the ground’, Sasak icaʔ ‘step on’. Dunnebier (1951) cross-references Mongondow itak to yitak and ritak. Since r and y vary freely in many Mongondow words, and the sequence yi- sometimes varies with i-, itak may simply be a phonologically advanced form of ritak, thus invalidating this comparison.

*icir creep, crawl

*icud budge, shift, move aside

*ida they, them

Note:   A number of languages in the Solomons and Vanuatu exhibit reflexes of earlier *kira. It is assumed that these derive from *ida (> *ira), with contamination from the other plural pronouns (*kita ‘1pl. incl.’, *kami ‘1pl. excl.’, *kamu ‘2pl.’). Tasiriki nira, Wailapa i-nira ‘they’ and similar forms in other languages of Vanuatu can be compared directly with *ni ida, but are assumed to be the result of an independent development.

*idal to fuss, complain; to rebel, cause a disturbance

*i(n)dan to praise, eulogize, honor someone

*idap hold a grudge

*idas₁ affine of Ego's generation

Note:   The meaning of this term is problematic. In PPh it apparently meant ‘spouse of sibling-in-law = sibling-in-law of spouse’, as this meaning is reflected both in Casiguran Dumagat of the northern Philippines, and in many of the MANO languages and Tiruray of the southern Philippines. Both Murut (Timugon) of northern Borneo and Mentawai, spoken west of central Sumatra, reflect a usage that is restricted to female speakers. If this restriction is part of the original specification of the term, *idas may have meant ‘husband's brother's wife’, but the confirming evidence remains tantalizingly elusive.

*idas₂ a strand, as of rope

Note:   Also Iban iraʔ ‘fibre, strand’.

*idau kind of snake

Note:   This comparison has benefitted from additional forms supplied by David Zorc.

*idi then, at that time

Note:   Probably identical to *i-di ‘that, there’.

*idus nasal mucus, snot

Note:   Also Malay (h)iŋus ‘snot; mucus from the nose’.

*ieq this

*ígat freshwater eel

*i(ŋ)git anger, resentment

Note:   With root *-git ‘anger, resentment’.

*ihap count, calculate

Note:   Also Ifugaw uyáp ‘act of counting, reckoning, computing’, Mongondow iab ‘count, calculate’. The semantic distinction between this item and *bilaŋ ‘count’ remains somewhat vague. What few indications we have suggest that *bilaŋ meant ‘count’ (viz. ‘recite numerals in sequence’), whereas *ihap meant ‘calculate’ (viz. ‘manipulate numerals to determine an amount’). Both of these clearly distinguishable senses are commonly conflated as ‘to count’ in English despite the availability of lexical resources for distinguishing them (‘count’ vs. ‘calculate’).

*ihi exclamation of surprise, etc.

*ijan when?

Note:   Also Paiwan ŋida ‘when?’, Hanunóo kapirá ‘when? when’, Maranao igira ‘when, while, during; whenever’, Sika rema pira ‘when?’, Tetun wen hira ‘when?’, Masiwang bot fila ‘when?’, Buruese beton pila ‘when?’, Buli ma-fis ‘when? (future)’, op-fis ‘when? (past)’, Dobuan ma-isa ‘when?’. Tagalog ka-ilán, Tboli k-ilon, Santa Catalina ka-ita ‘when?’ may point to an additional affixed form *ka-ijan. In any event, it is clear that the initial consonant of Tboli kilon is a fossilized affix rather than evidence for *qijan, since both Puyuma (Tamalakaw) a-ydan and Chamorro ŋa-iʔan contradict initial *q.

The reconstruction of *p-ijan is problematic. In many languages reflexes of *ijan are confused with reflexes of *pija ‘how much? how many?’ (see above), and it is possible that *p-ijan is a monomorphemic product of contamination. Whatever the case, *p- does not appear to be separable in any CMP language. Similarly, the great majority of OC languages (like Chamorro) reflect a form with prefixed *ŋa-, suggesting that POc *ŋaijan was monomorphemic. Contrary to this interpretation, a few languages in the Solomons (Mono-Alu, Santa Catalina) reflect *ijan without *ŋa-. Finally, both in Karo Batak and in Chamorro a reflex of *ijan is commonly used together with a morpheme nai. Attention to the glosses (Karo Batak nai ‘marker of past time’, Chamorro nai ‘where (relative), when (relative)’, however, suggests that this agreement is convergent, as apparently is that of Buli ma-fis ‘when (future)’, Dobuan ma-isa ‘when?’.

*ijuŋ nose

Note:   Also Kelabit idʰuŋ ‘nose’, Malay hidoŋ (expected **idoŋ) ‘organ of smell. Esp. of human noses, not animal snouts (monchoŋ)’, Old Javanese (h)iruŋ ‘nose’, Madurese eloŋ ‘nose’, Sasak eduŋ, eroŋ ‘nose’, Proto-Sangiric *idun ‘nose’, Tae' illoŋ ‘nose’, Palauan íis ‘nose’, isŋ-el ‘his or her nose’. Bimanese ilu ‘nose’, Wetan irna ‘nose’, Kamarian hiru ‘nose’, Lakalai ma-isu ‘nose’, Roviana isu (expected **isuŋu) ‘nose’, Central Maewo lisu- ‘nose’.

PWMP *ijuŋ-an probably had meanings other than that reconstructed here, as is suggested by Tagalog ilóŋ-an ‘big nosed’, Cebuano ilúŋ-an ‘having a nicely-shaped nose’, Manobo (Western Bukidnon) izuŋ-an ‘grow hard (rice grains)’, Javanese iruŋ-an ‘pipe linking blacksmith's bellows to a fire-box; nose-shaped latch on a folding door’. It is also possible that a verbal form *maŋ-ijuŋ was found in PWMP, as is suggested by Toba Batak maŋ-iguŋ-iguŋ ‘sniff at (of a dog)’, Sundanese ŋ-iruŋ ‘talk through the nose’. The reduplication in the foregoing Toba Batak form and in Samoan isu-isu ‘be inquisitive, put one's nose into other people's affairs’, on the other hand, probably is convergent. Finally, in POc, *ijuŋ apparently was extended metaphorically to capes of land (Nggela, Rotuman), and to prows of canoes (Lau, Maori, Hawaiian).