Chapter 7Nouns and possession
Noun morphology, the possessive (“belonging”) form, inalienable and alienable possession, and locative nouns.
The Sakhalin Ainu noun inflects for neither gender nor case, and number marking is optional (Chiri (1942: §§66–68); Chapter 8). Its one pervasive inflectional category is possession: nearly every noun opposes a bare conceptual form to a possessed form in -hV/-VhV, cise ‘house’ ~ cisehe ‘his/her house’. Kindaichi called the latter the “belonging form” (所属形); Chiri Mashiho 知里真志保, whose 1942 description remains the fullest, opposes the conceptual form (概念形) to person-marked forms built on a possessed stem (Chiri 1942: §§69–71). We write POSS for the suffix throughout. This chapter describes the formation and use of the possessed form, the syntax of possessor phrases, the inalienable–alienable divide, the locative noun subclass, and, in outline, derived nominals.
7.1 Conceptual and possessed form
Chiri introduces the opposition with sik ‘eye’ (Chiri's morphophonemic citation; the West-coast isolation form is sis by coda neutralisation, Chapter 4): the conceptual form names the bare concept, while the possessed form anchors the referent to a possessor, third person being zero-marked Chiri (1942: §70).
| form | gloss |
|---|---|
| sik | eye (conceptual form) |
| siki(hi) | his/her/their eye(s) |
| e-siki | your (sg) eyes |
| eci-siki | your (pl) eyes |
| ku-siki | my eyes |
| ci-siki | our (excl) eyes |
| an-siki | our (incl/indef) eyes |
7.1.1 Formation
Vowel-final stems copy their final vowel after an inserted h: sapa ‘head’ ~ sapa-ha, cise ~ cise-he, poo ‘child’ ~ poo-ho (Chiri 1942: §72). Stems in long ii/uu take -(y/w)ehe, imii ‘garment’ ~ imi-yehe Chiri (1942: §72), (Bugaeva 2022: ch. 9, §4.1), and in the corpus etu ‘beak’ ~ etu-wehe (example 5 of Chapter 6); short -i, -u and the glide codas are the domain of a Sakhalin innovation, the generalized -he, taken up in the next subsection. Unlike Hokkaidō, which opposes short tek-e to long tek-e-he, Sakhalin uses only the long shape (Bugaeva 2022: ch. 9, §4.1).
Consonant-final stems take -VhV. The descriptive tradition leaves the choice of vowel to the lexicon, with only loose tendencies by final consonant Chiri (1942: §72), (Murasaki 1979), but Chiri's later study of Ainu vowel harmony shows the choice to be largely rule-governed: tabulating some 160 noun stems, he found the suffix vowel agreeing in harmonic class with the last vowel of the stem, the classes being A (a, u), B (o) and C (e, i), and A and B vowels never combining within the word (Chiri 1952: 463–464). On this analysis nan ‘face’ ~ nan-uhu, ram ‘heart’ ~ ram-uhu and kotan ‘village’ ~ kotan-uhu select an A-class suffix vowel after A-class a/u, and sik ‘eye’ ~ sik-ihi a C-class vowel after C-class i — predicted, not memorized. What is genuinely lexical is the residue where harmony licenses more than one vowel and a stem (or a sense) settles on one: sar-a ‘its tail’ beside sar-i ‘its reed-field’ (Chiri 1952: 463). We treat such pairs as a short list of inflection-class diacritics, comparable to thematic vowels in declension systems elsewhere: the harmony rule assigns the class, and only the harmony-internal choices need be listed per noun. Dal Corso takes the underlying suffix vowel to be i and the first V of -VhV to be an epenthetic repair, citing free variation of the type kotanuhu ~ kotanhu in Piłsudski's corpus (Dal Corso 2024: 65, fn. 9); on that view the harmonic agreement states which quality the repair copies in, and the two analyses compose rather than compete.
Stems whose final stop has weakened to coda h resurface it before the suffix: teh ‘hand’ ~ tek-ihi, otoh ‘hair’ ~ otop-ihi, and mah ‘wife, woman’ ~ maci-hi, where restored t palatalizes before i (kuci from kut-i ‘his belt’, (Chiri 1942: §16)). The segmental alternations involved are treated in Chapter 5 and not repeated here; (2) shows the alternation live in one clause.
7.1.2 The generalized -he
Hokkaidō dialects treat nouns in final short -i and -u as special classes — ni-ye(he) ‘his tree’, with periphrasis usual for many u-stems — and give glide-coda nouns -e(he), haw-e(he). Sakhalin levels all of these under a single -he. In monosyllables the levelling is complete: nii ‘tree’ ~ ni-he, kuu ‘bow’ ~ ku-he Chiri (1952: 463), (Chiri 1942: §72), with the regular shortening of the lengthened monosyllable before the suffix (Chapter 5), and likewise for the glide codas, ay ‘arrow’ ~ ay-he, haw ‘voice’ ~ haw-he (Chiri 1942: §72). The corpus offers no counterexamples in these classes, and we state the rule as exceptionless there.
Polysyllabic -i/-u stems instead admit both shapes: Chiri records kampi-hi beside kampi-he ‘his letter’ (Chiri 1942: §72), and the conservative copying variant was still the form Murasaki's West-coast consultants produced, as (3) shows.
‘My teacup is not here.’
Murasaki 2025; West Sakhalin, Rayciska
The polysyllabic i-stem itanki keeps the copying -hi where generalized -he would also be licit.
The generalization has thus run to completion in just the short-stem classes — monosyllables and glide-finals — and remains variable in polysyllables: a lexically gradual, structurally bounded change rather than an across-the-board replacement. A prosodic signature travels with the morphology: where Hokkaidō shifts stress rightward onto the suffix, ru-é ‘its track’, Sakhalin keeps it on the stem, rú-he (Chiri 1942).
7.1.3 Choosing between the two forms
The conceptual form is used when the noun names a kind rather than an anchored referent: in compounding and incorporation (teh-kotoro ‘palm’, lit. hand-surface, (Bugaeva 2022: ch. 9, §4.3); (Chiri 1942: §76)), in vocatives (kakapo ‘elder sister!’, (Chiri 1942: §76)), and in classifying predications such as (4), where ‘wife’ denotes a role.
pirika maxnekux max ne koroči
‘They took beautiful women as wives.’
Chiri 1942: §76; Piłsudski 1912: 69; East Sakhalin
Chiri remarks that possessed maci here would wreck the sentence: ‘they took his wife’.
Once a referent is anchored in discourse, the possessed form tracks it. In the tale of the sticking tree-root, the axe is introduced in the conceptual form under the instrumental and immediately resumed in the possessed form, now ‘his axe’:
Tangiku, comparing parallel passages across the Asai corpus, finds exactly this division: first mention of an instrument may be conceptual or possessed, but the second mention within the scene is reliably possessed (Bugaeva 2022: ch. 9, §4.3). Where anchoring is irrelevant the conceptual form may stand even for owned items, so the choice is pragmatic, not automatic (Chiri 1942: §76).
7.2 The syntax of possession
Adnominal possession is head-marking and strictly head-final: the possessor NP stands before the possessed-form head, with no linking element — the type aynu cise-he ‘the man's house’. Pronominal possessor phrases are usually omitted, since the prefix identifies them (Bugaeva 2016: 104).
With no overt possessor and no prefix, the possessed form is third-person by default and is read from context, here ‘her husband’:
7.2.1 Person-marked possession
First- and second-person possessors are indexed on the possessed form by the same prefix set that marks transitive subjects: ku- 1SG, ci- 1PL exclusive, e- 2SG, eci- 2PL, an- indefinite Chiri (1942: §73), (Sakaguchi 2024: 19); Chapter 13. The corpus shows the set at work on nouns of every semantic type: ku-cip-ihi ‘my boat’ (9), ku-kema-ha ‘my legs’ (10), ku-nan-uhu ‘my face’ (Sakaguchi 2024: 19), eci-ona-ha ‘your father’ (11), an-ona-ha ‘my father’ (12).
tan ćíś ne ámpe kuáni ku-ćíphi nḗ
‘This boat is my boat.’
Piłsudski 1912: Nr. 16, lines 29–30; Sakaguchi 2024: 19; East Sakhalin
In first-person narration the indefinite set an- serves as narrative first person, on nouns as on verbs (Chapter 13):
Example (12) also shows the plural -hcin stacked outside the possessive suffix, a Sakhalin development without Hokkaidō counterpart: it pluralizes possessor or possessum according to context (Sakaguchi 2024: 110–115), (Sakaguchi 2020b); Chapter 8. In (13) it marks the several possessors of (severally possessed) heads.
7.3 Inalienable possession
Body parts, kin terms and part–whole nouns are the core of the possessive system: in connected speech they occur overwhelmingly in the possessed form, and a bare conceptual sapa or ona is largely confined to compounds, vocatives and kind-readings of the type in (4) Chiri (1942: §76), (Bugaeva 2022: ch. 9, §4.2). The corpus matches this distribution: hands, legs, heads, eyes, beaks, names, voices, fathers, mothers, sisters, wives, husbands and children appear as tek-ihi, kema-ha, sapa-ha, siki-hi, etu-wehe, ree-he, maci-hi, hoku-hu, poo-ho in example after example above. At the other pole, nouns for natural phenomena — cuh ‘sun’, ahto ‘rain’, atuy ‘sea’, wahka ‘water’ — normally stay conceptual, there being no possessor to anchor; Chiri notes that even these can be coerced into a possessed form when a god speaks of an-kanto-ho ‘our own sky’ (Chiri 1942: §76), and Murasaki's consultants allowed ku-atuy-ehe ‘my sea, the sea where I live’ (Bugaeva 2022: introduction). Inalienability is thus a strong statistical norm rather than a formal class boundary.
7.4 Alienable possession and the koro strategy
Here Sakhalin Ainu departs sharply from Hokkaidō. In Hokkaidō the suffixal possessed form is restricted to inalienables, and alienable possession requires the relative-clause periphrasis with koro ‘have’: a-kor mosir ‘our land, lit. the land we have’. In Sakhalin the suffixal construction has been generalized to nouns of all classes — houses (6), garments (7), boats (9), axes (5), villages and lands (an-mosir-ihi ‘our land’, (Murasaki 1979: 5)) — so that both inalienable and alienable relations are head-marked (Bugaeva 2016: 105); for the place of this appositive possessor-plus-possessed pattern within Ainu and the wider Pacific see (Bugaeva 2021). Even animate property is possessed suffixally: the ogre's hunting dog is its seta-ha (Murasaki 2001: text 35), an example Tangiku singles out as characteristically Sakhalin (Bugaeva 2022: ch. 9, §4.2).
The koro periphrasis nonetheless survives beside the suffixal pattern. Chiri describes it for possessees with personhood: an-koro mah ‘my wife’, nispa koro mah ‘the chief's wife’ (Chiri 1942: §76). The Sentoku letters use it for social relations and places, an=koro nispa ‘my dear sir’, e=koro kotan ‘your village’ (Tangiku & Ogihara 2001: Letter 1), and in the Asai corpus it alternates with the suffixal form within a single sentence:
‘They had children — they had both a boy and a girl.’
Murasaki 2001: text 39; West Sakhalin, Odasu (Asai Take)
Conceptual poo under koro for ‘to get children’; possessed poo-ho once the children are referential.
Where both strategies are available the difference is one of construal. Murasaki's consultant contrasted ku-kotan-uhu ‘my village’, said while in it, with ku-koro kotan, said of one's own village from elsewhere — and put ku-koro kotan in the mouth of a god (Murasaki 1979: 83–84), (Bugaeva 2022: ch. 9, §4.2). We read the suffixal form as marking a referent anchored to its possessor in the speech situation, and koro as predicating ownership or control; the corpus alternations in (14) and in the letters are consistent with this division.
7.5 Locative nouns
Locative (relational) nouns — onnayke ‘inside’, kaske ‘top’, empoke ‘underside’, sanke ‘vicinity’, okaake ‘place behind, aftermath’, etoko ‘place ahead’, uturuke ‘interval’, soyke ‘outside’ — form a noun subclass with their own syntax: they follow an NP directly (cise onnayke ‘the inside of the house’), and only they combine with the simple locative postpositions ta, wa, ene Chiri (1942: §79), (Bugaeva 2022: ch. 9, §5.1). Their case syntax belongs to Chapter 11; here we note only how they participate in the possessive opposition. Before a locative postposition the conceptual form appears, as in (15); as argument of an ordinary verb the possessed form is required, as in (16) and (17) (Bugaeva 2022: ch. 9, §5.4).
When a locative noun takes a personal possessor it selects the object prefixes, not the possessor set used by common nouns: en-ka ‘above me’ (Chiri 1942: §79), i-osmak-e ‘behind me’ in the narrative of Yamabe Yasunosuke 山辺安之助 (Sakaguchi 2024: 20). This object-like marking aligns the landmark argument of a locative noun with the object of a transitive verb, and we take it as the synchronic trace of their verb-like relationality.
7.6 Derived nominals
Nouns are derived from verbs chiefly with the formal nouns pe ‘thing, one’ / -p (poro-p ‘a big one’), hi ‘place, time, fact’, kur ‘person’ and usi ‘place’ (Chiri 1942: §§60–63); their syntax is treated with nominalization in Chapter 20. The nominaliser -p is given in Chiri's morphophonemic citation; in West-coast isolation it undergoes the regular coda-obstruent neutralisation of Chapter 4, as with sik ~ sis earlier in this chapter. The formal noun kur likewise appears in its morphophonemic shape, but its final r follows the separate coda-r alternation (West-coast r → h or an echo vowel, varying word by word) rather than the obstruent rule (Chapter 4). Distinctively Sakhalin is the extension of the possessive suffix itself to verbs: a clause may be nominalized by -hV on its predicate, as in sitayki-hi neampe ‘when (she) struck it’ in (2), e=ki-hi ‘how you are faring’ in the Sentoku letters (Tangiku & Ogihara 2001: Letter 1), or main-clause an-koro-ho ‘I have (a grandchild)’ noted already by Chiri (Chiri 1942: §77). Bugaeva analyses these V-POSS forms as genuine nominalizations and as the innovation that gave Sakhalin Ainu its distinctive non-finite clause types (Bugaeva 2016: 105–110). That they are nominal in syntax and not merely in shape is shown by their argument privileges: a V-POSS clause can stand as the subject of an evaluative or veridicality predicate, as in (18), where the nominalized clause is what is judged untrue.
‘What I said is untrue.’ (lit. ‘my saying is a lie’)
Murasaki 1979: 95; Bugaeva 2016: 108; West Sakhalin, Rayciska
The V-POSS construction and its uses return in Chapter 20 and Chapter 16.