Skip to content

Chapter 43The Alienable/Inalienable Split and Possessive Classification

Hokkaido Ainu's binary possessive split — obligatorily possessed inalienable nouns vs. periphrastically possessed alienable nouns — its semantic classes, gradience, and place on the alienability-typology map.

43.1 Two dimensions and three constructions

Hokkaido Ainu has no genitive case-particle, no possessive preposition or postposition, and no adjectival inflection for possession. Possession is instead expressed either by head-marking on the possessed noun itself or by a verb-based attributive clause using kor ‘have, possess’. The language is "predominantly head-marking, exhibiting … polysynthesis, and having a robust alienability opposition" Bugaeva (2021: 52).

Bugaeva, Nichols & Bickel identify two independent grammatical dimensions that should not be conflated Bugaeva (2021: 46–47, 58–60, 81–82). The first is possessibility — whether a noun's argument structure includes a possessor slot at all (optionally, obligatorily, or not at all). The second is the coding strategy — whether the relationship, when licensed, is expressed by head-marking on the possessed noun or by a verbal attributive clause with kor. These dimensions correlate with the traditional alienability distinction but the correlation is fixed grammatically rather than speaker-chosen: "a speaker does not decide to convey whether an item can be owned … the choice of a noun determines the possessive construction" Bugaeva (2021: 47).

The two grammatical dimensions of Hokkaido Ainu possession
dimensionvaluesHokkaido Ainu realisation
possessibilityoptionally possessible / obligatorily possessed / non-possessiblebody parts and kin = obligatorily possessed; artefacts, animals, places = optionally possessible; wild fauna and flora = typically non-possessible
coding strategyinalienable (head-marked) / alienable (verbal attributive)affiliative construction vs. kor construction

Three surface patterns cover "N₁'s N₂" in Hokkaido Ainu. The first two are possession constructions; the third is a modifier–head relation (quality, material, or kind) and is not possession Nakagawa (2024: 217–218); Satō (2008: 150–151):

#constructiontemplaterelationexample
1alienable (kor)N₁ kor N₂concepttransferable possessiona=kor cip 'my canoe'
2inalienable (affiliative)A-prefix=N₂affiliativenon-transferable possessionku=tek-ehe 'my hand'
3modifier (non-possessive)N₁concept N₂conceptquality / material / kindkane cise 'metal house'

The diagnostic that pattern (3) is not possession is person-substitutability: in patterns (1) and (2) the possessor N₁ can be replaced by a person-marking affix (ku=kor cip, ku=tek-ehe); in (3) it cannot — *ku=ru 'my footprint' is impossible, even though seta ru 'a dog's footprint' is well-formed Nakagawa (2024: 218). The two possession constructions are in complementary distribution per noun: the construction is determined by noun-class membership, not by the speaker's choice at utterance time Nakagawa (2024: 218); Bugaeva (2021: 47). Where that membership is gradient or construal-sensitive, the picture is more complex (§4 below). The morphophonology of the affiliative suffix is detailed in Chapter 45 (Morphophonology of the Affiliative Suffix and Echo/Copy Vowels), and the concept/affiliative distinction as a noun-class parameter in Chapter 44 (Concept Form vs. Affiliative (Possessed) Form).

43.2 The kor construction

In the alienable construction the possessed noun (N₂) remains in the concept form and is preceded by the transitive verb kor ‘have, possess’; the possessor (N₁ or a person-marking prefix) is the transitive subject of that verb: a=kor cip 'my canoe', a=kor makiri 'my knife', a=kor hayokpe 'my armour' Nakagawa (2024: 218). The semantics are those of alienable (transferable) possession: the possessed entity's existence is independent of the possessor's, and transfer to another person makes it theirs Nakagawa (2024: 218). The negative counterpart replaces kor with sak ‘lack’; predicative HAVE and BE schemas are treated in Chapter 98 (Predicative Possession: kor, Existential Possession, and isam), and the particle kor 'while' that is homophonous with the verb in Chapter 134 (Simultaneous kor 'while' and Overlapping-Event Clauses).

(1)
ku=kor 1SG.A=have
seta dog
hure be.red
toan DIST.DEM
kur person
kor have
cape cat
retar be.white

‘My dog is red; that person's cat is white.’

Hokkaidō Utari Kyōkai 1994; Hokkaido (dialect not further specified)

Minimal alienable pair: 1SG subject prefix ku= (first kor) vs. overt third-person possessor NP toan kur 'that person' (second kor). Both seta 'dog' and cape 'cat' remain in the concept form.

Bugaeva et al. analyse the kor construction as an appositive possessive with a verbal modifier: the attributive is a finite relative clause with kor whose object is gapped and coreferential with the head noun Bugaeva (2021: 55). The construction retains relative-clause properties: kor can carry perfective and plural morphology and supports relative-clause stacking Bugaeva (2021: 55–58). With a third-person possessor, kor loses overt person-marking and cliticises to the head noun — Anna katkemat kor-ninkari — a step toward head-marking that Bugaeva et al. characterise as ongoing grammaticalisation Bugaeva (2021: 59) ‹speculative›.

The kor construction also extends to belonging and social dominion, not only transferable ownership. When lexicalised in the compound N-kor-kur 'the one who holds N', it yields titles of authority such as kotan kor kur 'village chief' (lit. 'village-holding person') and the divine epithet nupuri kor kamuy 'the god who owns the mountain' Nakagawa (2024: 218–219); Bugaeva (2021: 56):

(2)
wa CONJ
kotan village
kor have
kur person
a=ne 4.A=COP
wa CONJ
an=an exist=4.S

‘I was the village chief.’

ILCAA Ainu materials 1976; Saru

Narrator uses narrative fourth person (a=, =an) for first-person reference throughout the text. kotan kor kur marks social dominion rather than transferable ownership of the village.

The set of nouns attested in the kor construction is lexically bounded. Drawing on Tamura's dictionary, Bugaeva et al. identify six classes of alienable nouns: people and kin (the largest class, 24 items including aynu, hapo, nispa), place and environment (11: kotan, mosir, pet), food and drink (7), tools and ritual objects (7: emus 'sword', ikor 'treasures'), domestic animals (including seta 'dog' and pewrep 'bear cub'), and human attributes (7: isoytak 'story', itak 'language', rametok 'bravery'). Wild animals and uncultivated plants are absent — nothing in the world is their owner Bugaeva (2021: 55–56); Tamura (2000: 88). A further affective use of kor with people nouns adds a nuance of endearment: a-kor pon menoko 'my dear young girl' Bugaeva (2021: 56).

43.3 The affiliative construction

In the inalienable construction the possessed noun takes the affiliative (possessed) form, derived from the concept form by a suffix that copies or echoes a vowel of the stem, and simultaneously carries a person-marking prefix drawn from the transitive-subject (A) set: ku= (1SG), e= (2SG), a= or an= (4th/indefinite), zero (3rd) Nakagawa (2024: 215–216); Satō (2008: 151, 154). Relational and spatial nouns deviate from this pattern by taking the objective (O) prefix set instead; they are treated in Chapter 52 (Relational and Spatial (Locative) Nouns and Their Possessed Forms). The A-prefix paradigm as a whole is set out in Chapter 62 (Architecture of the Personal-Affix System: The Four Persons and the S/A/O Paradigms). The possessor is thus encoded as a grammatical subject: Satō notes that e=paro 'your mouth' is literally 'you-subject mouth-in-relation', which reads oddly when translated word-for-word but reflects exactly the head-marking logic of the language Satō (2008: 150).

(3)
ku=kema-ha 1SG.POSS=leg-POSS
ka even
arka hurt
ku=tek-ehe 1SG.POSS=hand-POSS
ka even
arka hurt

‘My leg hurts too, and my hand hurts too.’

Hokkaidō Utari Kyōkai 1994; Hokkaido (dialect not further specified)

kema 'leg' is vowel-final (type I) and takes the affiliative suffix -ha; tek 'hand' is type III-2 and takes -ehe. Both carry the 1SG possessive prefix ku=. The possessed body-part NP is the single argument of the intransitive verb arka 'hurt'.

The semantic classes of nouns that take the affiliative construction span five groups Nakagawa (2024: 219–221):

  • Body parts and normally worn or carried items. sik ‘eye’sik-ihi; par ‘mouth’par-o (irregular); tek ‘hand’tek-ehe; kema ‘leg, foot’kema-ha. Clothing normally worn is treated analogously (a=tesma-ha 'my snowshoes') and carried tools in some contexts (a=tasiro-ho 'my mountain-knife') Nakagawa (2024: 219).
  • Part–whole relations. at sinta atu-hu 'cradle cord'; rur sayo ruri-hi 'gruel broth' Nakagawa (2024: 219–220).
  • Kinship. A subset of kin terms takes the affiliative construction and obligatory person-marking; the full paradigm, the address/reference contrast, and dialect variation in kin-term assignment are set out in Chapter 48 (Kinship and Honorific Possession) Nakagawa (2024: 221–225); Satō (2008: 151).
  • Material-to-product. ohaw ihawe-he 'the taste/essence of the soup' Nakagawa (2024: 220).
  • Miscellaneous relational. atay-e-he 'price'; hura-ha 'smell'; tokuy-e-he 'best friend'; asuru-hu 'rumour' Nakagawa (2024: 220).

The sensory nouns haw ‘voice’, sir ‘appearance, situation’, hum ‘sound, feeling’, and ru ‘track, path’ participate in the affiliative construction as ordinary possessed nouns before their possessed forms grammaticalised as evidential formal nouns (treated in Chapter 120 (The Nominalization-plus-Copula Evidential Schema) Nakagawa (2024: 258–259)). The possessive use of hawe 'voice' is attested in ILCAA oral-literature texts ILCAA Ainu materials (1976). Community analysis of the form e=hawe aste 'raise your voice' similarly parses e=hawe as the possessed noun '2SG.POSS=voice' under the verb aste, confirming that hawe remains a productive possessed form alongside its evidential use (nukopoli, aynu-corpora Discord, 2025-12-03; aynu-corpora Discord (2023–2026)) ‹corpus-confirmed›.

The collective noun utar ‘companions, kin, community members’ provides a further illustrative case. A community-proposed analysis holds that utari(hi) is the affiliative form of utar, narrowing the sense toward '(someone's) companions / kin' in the same way the affiliative form generally specifies a relational reading against the more general concept form (thegodofneet, aynu-corpora Discord, 2025-09-17; aynu-corpora Discord (2023–2026)). This is consistent with Nakagawa's treatment of k=utari as a possessed kin-group NP Nakagawa (2024: 251–252) ‹contested›.

A related set of affiliative-type prefixese- 'its tip/head of' and o- 'its end/base of' — prefix to verbs and behave analogously to affiliative nouns in requiring a possessive antecedent Kobayashi (2025: 171). These are homophonous with the applicative prefixes e-/o- that increase valency by one; Kobayashi distinguishes them by valency neutrality and by restriction to unaccusative intransitive hosts Kobayashi (2025: 171) ‹contested›. The applicative prefixes and the overlap with relational nouns are taken up in Chapter 52 (Relational and Spatial (Locative) Nouns and Their Possessed Forms).

43.4 Class membership: gradient and construal-sensitive

The core of the noun-class split is unambiguous. Body parts, genealogical kin terms, and inalienable relational concepts are obligatorily possessed and take the affiliative construction; artefacts, domestic animals, and transferable resources take kor and cannot take a person-marking prefix directly Bugaeva (2021: 52, 55). Satō's diagnostic is exact: 'money' (kane) takes only ku=kor kane; the form *ku=kane is impossible. 'Mouth' (par) takes only e=paro; *e=kor par is impossible Satō (2008: 150–151). The two-construction distinction is documented in all major descriptive and pedagogical accounts Nakagawa (2024: 217–221); Satō (2008: 149–156); Ijäs (2023).

Obligatorily-possessed nouns are bound: they almost never appear without possessive morphology. When cited in isolation, speakers give the 3rd-person (zero-prefix) affiliative form — sik-ihi 'eye(s), his/her eye', never bare sik. In riddles, where no specific possessor is identified, the 4th-person (indefinite) A-prefix takes over: a=sik-ihi 'people's eyes' Bugaeva (2021: 53). These bound-noun properties, and the difference between nominal subclasses with and without obligatory possession, are developed in Chapter 34 (Nominal Subclasses and Obligatorily-Possessed (Bound) Nouns).

When an obligatorily-possessed noun is incorporated with a non-coreferential possessor, the antipassive prefix i- encodes the unspecified possessor and allows the bare noun root to incorporate: i-sapa-kar 'cut someone's head', i-rekut-tuye 'cut someone's throat' Bugaeva (2021: 53–54). Hokkaido Ainu thus has two distinct de-specifying strategies: the indefinite A-prefix a=, which retains possessive morphology and keeps the noun as a free NP, and the antipassive i-, which suppresses the possessor and licences incorporation. These are treated in Chapter 89 (Subject Incorporation and Possessor-Stranding) Bugaeva (2021: 54, 81).

Several nouns fall outside the clean bipartition. The noun kotan ‘village’ takes both constructions approximately equally: a=kor kotan ~ a=kotanu. The noun mosir ‘world, land, island’ takes only the kor construction in prose (*a=mosirihi is unattested, though mosiri appears in verse). The noun cise ‘house’ is dialect-split: Saru, Chitose, and Horobetsu prose uses only kor for 'so-and-so's house', but verse and other dialects allow the affiliative cisehe; the compound uncise 'home, dwelling-place' takes the affiliative exclusively across dialects Nakagawa (2024: 220–221).

The clearest evidence for construal-sensitivity in class assignment comes from Nakagawa's discussion of the trap-bow (ku). In the Ainu worldview, the prey animal chooses whose trap-bow to fall to, so the bow always pre-belongs to a specific person: it is construed as inalienable and takes the affiliative form Nakagawa (2024: 229):

(4)
tan this
pe thing
a=po-ho 4.A=son-POSS
ku-wehe bow-POSS
ne COP

‘This is my son's (bow).’

Nakagawa 2024; Hokkaido (dialect not further specified)

Inalienable (affiliative) construction for the trap-bow. ku-wehe is the affiliative of ku 'bow'; poho is the affiliative of po 'son'. The possessor a=poho 'my son' is itself in the affiliative form.

The same bow, framed as an ordinary transferable object, takes the kor construction Nakagawa (2024: 229):

(5)
tan this
ku bow
a=po-ho 4.A=son-POSS
kor have
ku bow
ne COP

‘This is a bow that my son has.’

Nakagawa 2024; Hokkaido (dialect not further specified)

Alienable (kor) construction for the same referent. The concept form ku appears twice: as the topic NP and as the gapped object of kor. Nakagawa uses the pair to show that the construction choice is construal-sensitive, not fixed by the referent.

The inalienable frame foregrounds that the bow has a pre-assigned owner in a cosmological sense; the alienable frame treats it as a tool in ordinary human hands. This pair is the strongest available evidence that the noun-class split, while grammatical in its effects, is not wholly referent-determined ‹contested›. A parallel diachronic observation runs in the same direction: Bugaeva et al. note that loanwords such as kosonde 'kimono' enter as non-possessible and take kor, suggesting that unfamiliarity is the default entry point for new nouns and that possessibility can be acquired over time Bugaeva (2021: 57, 76–77) ‹speculative›.

43.5 Analytical perspectives

The affiliative trigger has been debated in terms of both what the formal condition is and what typological category it instantiates.

The oldest question in the descriptive literature is whether the affiliative form is licensed by the specificity of the possessor or by a strictly grammatical condition. Tamura held that N₂ takes the affiliative only when N₁ is specific: the contrast between kamuy rus-ihi 'the specific bear's pelt' and kamuy rus 'bear-pelt in general' was the key datum, the latter approaching the status of a compound Nakagawa (2024: 231). Satō objected that the affiliative appears with non-specific possessors: aynu opitta hon-ihi kor wa okay pe ne 'all humans have a belly' — where aynu opitta 'all people' is maximally non-specific — carries the affiliative form hon-ihi Satō (2008: 155); Nakagawa (2024: 231) ‹contested›. Nakagawa's analysis shifts the frame entirely: the affiliative form appears wherever the noun carries a person-marking A-prefix or occupies the modifier slot of a relative-clause structure, regardless of possessor specificity Nakagawa (2024: 230–233). Three constructions support this: the indefinite-person affiliative (a=teke munin pe ne na 'one's hand will rot', a general warning); affiliatives inside relative clauses (the modifier 'with a long tail' requires sara-ha tanne, not *sar tanne, because the modifier slot is a relative-clause position); and affiliatives triggered by quantifiers usa 'various' and nep 'whatever' Nakagawa (2024: 231–233) ‹contested›.

Satō's personal-form (人称形) analysis, developed in Chapter 44 (Concept Form vs. Affiliative (Possessed) Form), addresses a different aspect. Once a person-marking prefix combines with the affiliative form to create a personal form, the result is a highly autonomous NP: Ø=sik-ihi alone — without any overt possessor NP — is a complete noun phrase 'his/her eye'. At the same time, the possessor NP and its personal form need not be adjacent: the noun phrase menoko … anakne … kaskamuye 'the woman's guardian spirit' has the topic particle anakne intervening. The personal form signals the presence of a possessor somewhere in context rather than forming a tight N-N genitive Satō (2008: 155–156). Bugaeva et al. converge on the same point: the possessed suffix "marks a noun as requiring a possessor (which is then indexed with a possessive prefix)" and "the possessor N₁ can easily be omitted without ambiguity as the possessed suffix signals its presence" Bugaeva (2021: 52).

Bugaeva et al.'s typological framing, drawing on (Nichols 1986), places the Hokkaido system within the Circum-Pacific appositive possessive type but distinguishes it by the fact that the modifier is a finite verbal clause (kor) rather than a nominal appositive Bugaeva (2021: 58–60, 80–82). The two constructions together give the language a mixed locus-of-marking profile: inalienable possession is head-marked on the possessed noun (A-prefix + affiliative suffix), while alienable possession is encoded as a verbal modifier. The adnominal double-marking and appositive constructions that arise at the interface of these two strategies are examined in Chapter 46 (Adnominal Possession: Double Marking and the Appositive Construction); the external-possession and possessor-raising phenomena are treated in Chapter 47 (External Possession and Possessor Raising).

Huber's subjectivity analysis applies Ikegami's HAVE/BE-language typology and finds that Ainu patterns with English (a HAVE-type language where the possessor is foregrounded as subject) rather than Japanese (a BE-type language where the possessed entity is topicalised) Huber (2025: 254–257). Even in the inalienable construction the possessor is overtly indexed by the A-prefix — ku=tek-ehe ku=kikkik 'I clapped my hands' — unlike the dropped possessor of Japanese 「手を叩いた」. Huber proposes treating inalienability as a subjectivity index and the possession-schema inventory as a gradient scale rather than a binary; this proposal is explicitly framed as preliminary in her paper Huber (2025: 257) ‹speculative›.

References cited in this chapter

aynu-corpora Discord (2023–2026) ·Bugaeva (2021) ·Hokkaidō Utari Kyōkai (1994) ·Huber (2025) ·Ijäs (2023) ·ILCAA Ainu materials (1976) ·Kobayashi (2025) ·Nakagawa (2024) ·Nichols (1986) ·Satō (2008) ·Tamura (2000)