Chapter 34Nominal Subclasses and Obligatorily-Possessed (Bound) Nouns
The internal taxonomy of nouns — concrete vs abstract, common vs proper, and the class of 'incomplete'/bound nouns that cannot stand without a possessor.
34.1 The four noun subclasses
A noun is defined as a word that can form an NP and serve as a verb's subject or object Nakagawa (2024: 214). Within the class, four subclasses are distinguished by two crossing diagnostics: whether the form can stand alone as an NP argument, and whether it can serve as the head of a relative-clause structure Nakagawa (2024: 215–216); Nakagawa (2022: 480–482). Crossing these two properties against the type of person affix the form takes and its status as a locative expression (⟨place⟩) yields a six-row grid.
| subclass | form | stands alone as argument? | can head a relative clause? | person affix type | functions as ⟨place⟩? |
|---|---|---|---|---|---|
| common noun (普通名詞) | concept form (概念形) | ○ | ○ | — | no |
| common noun | affiliative/possessed form (所属形) | ○ | ○ | nominative (possessive) | no |
| locative noun (位置名詞) | basic form | ✗ (Type 1) / ○ (Type 2) | ✗ | objective | yes |
| locative noun | long form | ○ | ✗ | objective | yes |
| pronoun (代名詞) | — | ○ | ✗ | — | some (via formal noun hi) |
| formal noun (形式名詞) | — | ✗ | ○ | — | some (hi, uske, usi…) |
The rows fall into an autonomy hierarchy. Common nouns in the concept form are the most independent; common nouns in the affiliative form presuppose a possessor relation but can still stand as NP arguments. Locative nouns are always ⟨place⟩ and take objective rather than nominative person affixes; the Type-1 basic form is the least autonomous among the locative forms, requiring an antecedent NP. Pronouns can stand alone but cannot be relativised. Formal nouns are the least autonomous of all: they require a preceding modifier or clause and cannot appear as bare NP arguments Nakagawa (2024: 215–216).
Pronouns and formal nouns receive full treatments in separate chapters: Chapter 49 (Personal, Reflexive, and Emphatic Pronouns) and Chapter 35 (The Formal / Defective Noun Set (ruwe, hawe, siri, humi, hi, pe, kur)). Locative nouns are introduced in this chapter and developed in detail in Chapter 52 (Relational and Spatial (Locative) Nouns and Their Possessed Forms).
34.2 Common nouns: concept form and affiliative form
Every common noun has two forms. The concept form (概念形) presents the referent without anchoring it relationally to anything else. The affiliative form (所属形) presents it in relation to a possessor or reference point, and it is the form that carries a personal prefix indexing that possessor. Kindaichi & Chiri used the terms 抽象形 (abstract form) and 具体形 (concrete form) for the same distinction Kindaichi & Chiri (1936: 16); the current terminology, following Nakagawa and the Bugaeva Handbook, replaces those labels without changing the underlying analysis Nakagawa (2024: 217); Nakagawa (2022: 484).
Three possession constructions are built on this opposition Nakagawa (2024: 217–219):
| # | construction | N₂ form | relation expressed | example |
|---|---|---|---|---|
| (1) | N₁ kor N₂ / affix=kor N₂ | concept form | alienable | a=kor cip 'my/his boat' |
| (2) | N₁ N₂-affix / affix=N₂-affix | affiliative form | inalienable | ku=tekehe 'my hand' |
| (3) | N₁ N₂ (juxtaposition, no affix) | concept form | attributive (not possession) | menoko tek 'a woman's hand' |
‘a woman's hand (attributive, not possessed)’
constructed example
Attributive juxtaposition, construction (3): menoko 'woman' specifies the attribute of tek 'hand' but is not its possessor in the grammatical sense. No person prefix can substitute for menoko here: *ku=tek is ungrammatical as a possessive.
The decisive diagnostic distinguishing (2) from (3) is the person-prefix substitution test: the first NP of an inalienable construction can be replaced by a person prefix (ku=sikihi 'my eye' — concept form sik, affiliative sikihi), whereas the first noun of an attributive construction cannot (*ku=rus for 'bear-pelt'; only kamuy rus or kamuy rusihi 'that specific bear's pelt' is possible) Nakagawa (2024: 218). The morphophonology of affiliative-form formation — the vowel-copy and echo-vowel suffixes — is treated in Chapter 45 (Morphophonology of the Affiliative Suffix and Echo/Copy Vowels). The full semantics of the concept/affiliative opposition, including its non-possessive (anaphoric and definiteness) uses, are developed in Chapter 44 (Concept Form vs. Affiliative (Possessed) Form).
34.3 Inalienable nouns and the bound class
The inalienable construction (construction 2) covers several semantic fields. Nakagawa identifies five main classes that favour the affiliative form when possessed: (a) body parts and items typically worn or carried (ku=tekehe 'my hand', a=tesmaha 'my snowshoes'); (b) part-of-whole nouns (op nicihi 'spear-shaft', sayo rurihi 'gruel broth'); (c) a subset of kinship terms (see below); (d) material-to-product nouns (ikaray ohawe 'nettle stew', top cise 'bamboo house'); and (e) a miscellaneous set including ataye 'price', hura ‘smell’ (affiliative huraha), and kotanu 'village' Nakagawa (2024: 219–220).
Within the broader inalienable class, a distinction can be drawn by degree of boundness. Common nouns in the body-part and part-whole classes retain an independent concept form that can appear in generic or classifier-like contexts: tek 'hand' can appear as a bare NP argument. By contrast, the Type-1 basic forms of locative nouns (§ Chapter 52 (Relational and Spatial (Locative) Nouns and Their Possessed Forms)) cannot appear at all as NP arguments without an immediately preceding antecedent, and formal nouns cannot appear without a preceding clause or prenominal modifier. These two groups represent the strictly bound end of a gradient from "tends to be possessed" to "cannot appear without an antecedent" Nakagawa (2024: 238–240, 257).
34.3.1 Conditions on the affiliative form
A long-standing descriptive question concerns what triggers the affiliative form specifically. Tamura (1988a: 27–28, as reported by Nakagawa 2024: 228) proposed a specificity condition: N₂ appears in the affiliative form only when N₁ is specific and individuated (kamuy rusihi 'that specific bear's pelt' vs kamuy rus 'bear-pelt in general', which is almost a compound) ‹contested›. Satō objects that the affiliative does appear with non-specific possessors — aynu opitta honihi kor 'all humans have a belly' — which would be excluded by a strict specificity condition Nakagawa (2024: 228–229) ‹contested›. Nakagawa's 2024 account resolves the tension by placing the conditioning in morphosyntax rather than semantics: whenever a person prefix is present, the affiliative form is required; specificity then emerges as a tendency rather than a categorical trigger, because inalienable possession of an individuated entity is simply the most common context for person-prefix use Nakagawa (2024: 228–230).
Nakagawa further notes that the sharp "has/lacks affiliative form" dichotomy, sometimes presented as a fixed lexical property, is overstated: given an appropriate relational context, most common nouns can in principle take an affiliative form Nakagawa (2024: 227). The noun utar ‘plural marker / people / family’ illustrates the gradient: as a formal-noun plural marker utar is uninflected, but as a common noun 'relatives, companions' it has an affiliative form utari(hi). Community discussion has noted that, because it is hard to conceive of 'companions not belonging to anyone', this common noun surfaces almost exclusively in its affiliative form as 'N's family/companions' aynu-corpora Discord (2023–2026) (thegodofneet, 2025-09-17) ‹contested›, extending Nakagawa's account of the same ambiguity at Nakagawa (2024: 251–252).
34.3.2 Kinship terms: two grammatical classes
Kinship terms split into two sets with different possession constructions, and the assignment is lexically specified. Terms using the kor-construction (concept form): ku=kor hapo 'my mother', ku=kor mici 'my father'. Terms using the affiliative construction: ku=turesi 'my younger sister (sister's viewpoint)', ku=yupihi 'my elder brother' Nakagawa (2024: 221–222). Within Hokkaido there is considerable dialect variation in the assignment of individual kinship terms to one set or the other Nakagawa (2024: 219–222).
A further division concerns address vs reference. Several kinship terms — among them ona 'father', unu 'mother', yup 'elder brother', and sa 'elder sister' — are restricted to third-person reference and cannot be used in direct address; each has an address counterpart. The restriction correlates with the grammatical pattern: only the affiliative form combines with a person prefix, so a term that cannot be addressed directly tends not to appear with first- or second-person possessor marking Nakagawa (2024: 221–223). Kinship possession in full, including suppletive address forms and the dialect-by-dialect distributions, is treated in Chapter 48 (Kinship and Honorific Possession).
34.3.3 Membership, gradience, and dialect variation
Inalienable-class membership is partly dialect-specific and partly lexeme-specific. kotan ‘village, settlement’ takes both the kor-construction and the affiliative construction equally in Saru; mosir ‘world, land’ uses only the kor-construction (a=kor mosir; *a=mosirihi is ungrammatical) Nakagawa (2024: 220). These lexeme-specific idiosyncrasies are part of the lexical entry and are not predictable from semantics alone. Refsing notes comparable fluctuation in Shizunai, where some body-part stems allow either construction Refsing (1986: 74–76).
The full alienable/inalienable distinction, adnominal possession constructions, and the interaction of possession with nominal number and person marking are developed in Chapter 43 (The Alienable/Inalienable Split and Possessive Classification), Chapter 44 (Concept Form vs. Affiliative (Possessed) Form), and Chapter 46 (Adnominal Possession: Double Marking and the Appositive Construction).
34.4 Locational and relational nouns
Locative nouns form a small, closed subclass expressing spatial and temporal relations ('above', 'below', 'in front of', 'beside', 'inside'). Four properties set them apart from common nouns Nakagawa (2024: 238–240); Nakagawa (2022: 490–491):
- The basic form cannot stand alone as an NP. Type-1 locative nouns require an immediately preceding antecedent NP (the spatial reference point). Without it, *ka wa ran 'came down from above' with bare ka is ungrammatical; only the long form can head a clause independently (§ below).
- Long-form derivation differs from the common-noun affiliative. Common nouns use a vowel-copy suffix (the -hV mechanism) predictable from the stem's final vowel; locative long forms use a distinct suffix set — -ke, -V, -si — that is lexically specified per item and follows different morphophonological patterns. The overlap between the two systems may be historical: Nakagawa hypothesises that common-noun type-③-2 morphology was borrowed from the locative system Nakagawa (2024: 226, 247) ‹speculative›.
- Both forms take the objective person affix. Common nouns in the affiliative construction take the nominative (possessive) prefix; locative nouns take the objective prefix: en=sam un 'to beside me' (1SG.O prefix en= on sam ‘beside’) vs ku=tekehe 'my hand' (1SG.A/possessive prefix ku= on tek ‘hand’) Nakagawa (2024: 241).
- Both forms always function as ⟨place⟩. Locative nouns can be directly followed by the locative particle ta: kasi ta 'on top of it'. A common noun cannot do this: *nupuri ta is ungrammatical; 'at the mountain' requires an intervening locative noun, nupuri or ta ('at [the inside of] the mountain') Nakagawa (2024: 216, 241).
34.4.1 Type-1 and Type-2 locative nouns
Nakagawa further divides locative nouns into two types Nakagawa (2024: 241–246). Type-1 nouns require an immediately preceding antecedent NP in the basic form; in the long form they can head a clause alone. Type-2 nouns can stand in the basic form without naming a reference point, making them syntactically closer to common nouns, though they retain the locative-noun properties of ⟨place⟩ status and objective-affix agreement.
| basic form | long form | meaning | type |
|---|---|---|---|
| ka | kasi(ke)(he) | above (in contact) | 1 |
| enka | enkasi(ke)(he) | above (at a distance) | 1 |
| kurka | kurkasi(ke)(he) | over a spread-out surface | 1 |
| corpok | corpok(i)ke(he) | below (root pok) | 1 |
| etok | etokoho | front of a moving/linear thing; temporal 'before' | 1 |
| kotca | kotcake(he) | front of a stationary thing | 1 |
| oka | okake(he) | back of a moving/linear thing; temporal 'after' | 1 |
| osmak | osmake(he) | back of a stationary thing | 1 |
| sam | samake(he) | beside | 1 |
| utur | uturu(hu) / uturke(he) | between | 1 |
| or | oro / orke(he) | inside, at the location of | 1 |
| oske | oskehe | inside (general) | 1 |
| tum | tumuke(he); contracted tumke ‹corpus-confirmed› | inside a homogeneous mass | 1 |
| noski | noskie / noskike(he) | middle | 1 |
| soy | — | outside | 2 |
| rik | — | high place | 2 |
| kim | — | mountain-ward | 2 |
| pis | — | beach-ward | 2 |
| ya | — | shore, land-side | 2 |
| mak | — | inner / temporal-after | 2 |
The distinction between etok (front of a moving entity or linear path; temporal 'before') and kotca (front of a stationary entity) carries real semantic weight: i=etok omare 'placed in front of me [as a moving being]' means the referent was set ahead on the path, not merely in the visual field Nakagawa (2024: 243). Similarly, oka (back of a moving entity; temporal 'after') and osmak (back of a stationary entity) form a paired contrast.
The locative noun or ‘inside, at the location of’ has developed an extended use in several Hokkaido dialects, functioning as a broad oblique marker; its grammaticalization in Chitose is documented in Satō (2021).
‘I stood on top of it.’
Nakagawa 2024: 239; Chitose, Nakagawa ex. 278
The long form kasi (ka + -si, extendable to kasike(he)) of Type-1 locative ka heads the clause alone; the basic form *ka could not appear here without a preceding antecedent NP. The fourth-person suffix =an carries narrative-first-person reference.
‘It has come to settle beside me.’
Nakagawa 2024: 241; Chitose, Nakagawa ex. 281
Objective prefix en= (1SG.O) attaches directly to the Type-1 locative sam 'beside'. The allative particle un follows, as sam is ⟨place⟩. Contrast the nominative/possessive prefix ku= that would appear on a common-noun affiliative.
‘It entered the river and died.’
Nakagawa 2024: 245; Hokkaido (dialect not further specified), Nakagawa ex. 299, Samani (様似)
Type-1 locative oske 'inside (hollow space)' immediately follows its antecedent pet 'river', as required for the basic form. The basic form cannot head a clause alone.
‘When I went to the mountain…’
Nakagawa 2024: 246; Chitose, Nakagawa ex. 300
Type-2 locative kim 'mountain-ward' heads the clause in basic form with no antecedent. The common noun nupuri 'mountain' is not ⟨place⟩ and requires nupuri or ta to express the same location.
34.4.2 The Type-2 classification controversy
Some researchers classify Type-2 locative nouns — kim, soy, rik, and the like — as common nouns on the grounds that they appear independently in the basic form, a property otherwise confined to common nouns. Nakagawa retains them as locative nouns on two grounds: their long-form morphology follows the locative-noun pattern rather than the common-noun affiliative system, and their obligatory ⟨place⟩ status distinguishes them from any common noun, which would require a locative noun to mediate its use with ta Nakagawa (2024: 245–246) ‹contested›.
One community observation proposes that the long forms of locative nouns carry a definiteness or high-specificity marking, which might explain why first- and second-person pronominal reference is expressed on the long form rather than the basic form aynu-corpora Discord (2023–2026) (antitwilight, 2024-11-17) ‹speculative›. This hypothesis is not attested in the primary-source grammars; it awaits cross-dialectal elicitation.
The directional adverbs derived from Type-2 forms (ekimne 'toward the mountains', okimne 'from the mountains'; esoyne 'outward') and the full locative-noun possessive paradigm are treated in Chapter 52 (Relational and Spatial (Locative) Nouns and Their Possessed Forms).
34.5 Common vs proper nouns
Proper nouns form no grammatically distinct class in Hokkaido Ainu: they are an open subset of common nouns, distinguished by reference type rather than by any morphological marker, case distinction, or definiteness operator Nakagawa (2024: 233); Nakagawa (2022: 483–484). Many place names are transparently compositional: Poropet 'big river' (poro + pet ‘river’), Siretoko 'tip of the land' (sir ‘land, appearance’ + etoko) Nakagawa (2024: 233).
One property that distinguishes proper nouns from common nouns in practice: Hokkaido personal and place names do not appear in affiliative forms under normal conditions, unlike many common nouns Nakagawa (2024: 233). Sakhalin Ainu permits affiliative forms of personal names — ku=Harumihi 'my Harumi' is attested Murasaki (1979) — but this pattern is absent from the Hokkaido data.
Personal-name avoidance, the practice of late and deliberately unflattering naming to deflect malevolent attention, and the classification of kamuy names as proper vs common reflect cultural conventions rather than grammatical rules Nakagawa (2024: 234–236).
34.6 Pronouns and formal nouns in the subclass system
Personal pronouns — kani 'I', eani 'you (2SG)', coka 'we (exclusive)' — share the common-noun property of standing alone as NP arguments but differ from common nouns in that they cannot head relative clauses: there is no grammatical Ainu equivalent of English 'the I who is sleeping' Nakagawa (2022: 482). Their paradigms, dialect variants, and the fourth-person forms are treated in Chapter 49 (Personal, Reflexive, and Emphatic Pronouns).
Formal nouns — also glossed as noun particles (名助詞) — occupy the bottom of the autonomy hierarchy. Unable to appear as bare NP arguments, they nominalise a preceding clause or prenominal element into an NP head Nakagawa (2024: 257–258). The class includes p ‘thing (after vowel)’ ~ pe ‘thing (after consonant)’, hi ‘thing, matter, time, place’, katu ‘state, manner’, kur ‘person’, utar ‘plural marker / people’, and the evidential nominalizers ruwe ‘trace, track (possessed form: the trace of —)’, siri ‘appearance (possessed form: the appearance of —)’, hawe ‘voice (possessed form: the voice of —)’, and humi ‘sound, sensation (possessed form: the sensation of —)’, whose possessed forms build the evidential paradigm Nakagawa (2024: 258). The class is defined and its membership criteria set out in Chapter 35 (The Formal / Defective Noun Set (ruwe, hawe, siri, humi, hi, pe, kur)); the nominalization function of p/pe and hi in relative-clause and complement constructions is treated in Chapter 101 (The Nominalizers -p/-pe and -i/-hi (Participant, Event, Place, Fact)); the four evidential nominalizers are treated in Chapter 120 (The Nominalization-plus-Copula Evidential Schema), Chapter 121 (ruwe ne — the Inferential / Visual-Trace Evidential), Chapter 122 (siri ne — the Direct-Perception / Situational Evidential), Chapter 123 (hawe ne — the Reportative / Hearsay Evidential), and Chapter 124 (humi ne — the Non-Visual Sensory Evidential).
References cited in this chapter
aynu-corpora Discord (2023–2026) ·Kindaichi & Chiri (1936) ·Murasaki (1979) ·Nakagawa (2022) ·Nakagawa (2024) ·Refsing (1986) ·Satō (2021)