Chapter 35The Formal / Defective Noun Set (ruwe, hawe, siri, humi, hi, pe, kur)

35.1 Overview and terminology

Among the four noun subclasses of Hokkaido Ainu, the formal noun (形式名詞 keishiki meishi) occupies the lowest position in the independence hierarchy: none of its members can stand alone as a predicate argument. Each must be preceded by a verb phrase, a prenominal, or — in one exceptional case — a bare noun, from which it draws its referential content Nakagawa (2024: 257); Nakagawa (2022: 480–482). The diagnostic that separates formal nouns from the three other subclasses (common, locative, and pronominal nouns) is set out in Chapter 33 (The Word-Class Inventory and the Noun/Verb Bipartition). Nakagawa also calls them noun particles (名助詞 meijoshi), reflecting their particle-like attachment to the preceding element Nakagawa (2024: 257, 295).

The terminology used for this class is unsettled ‹contested›. Satō (2008: 174–178) divides what Nakagawa treats as a single class into two: nominalizers (名詞化辞 meishikaji) head content clauses with full argument structure, while formal nouns proper head relative clauses from which one argument is missing. Both are distinct noun subclasses in his account. Nakagawa keeps the single class label and instead distinguishes the role a formal noun plays clause by clause: as relativizer it heads a clause with a missing argument; as complementizer (補文標識 hōbun hyōshiki) it heads a clause with full argument structure, the diagnostic being arity — a clause summing to arity zero takes a complementizer, one with a net argument gap heads a relative construction Nakagawa (2024: 503–504). Refsing (1986: 259–264) uses nominalizer more broadly still, grouping items that other analyses class as mood-morphemes, complementizers, and relative-clause heads within a four-type nominalization typology ‹contested›. The full accounts of relativization and complementation are in Chapter 101 (The Nominalizers -p/-pe and -i/-hi (Participant, Event, Place, Fact)) and Chapter 102 (Lexical-Head and Formal-Noun Nominalization (kur, uske, ruwe/hawe/siri/humi)); the evidential constructions are in Chapter 120 (The Nominalization-plus-Copula Evidential Schema).

35.2 The inventory

Nakagawa (2024: 257–258) lists two sets. The general set covers a range of relational senses; the sensory-evidential set (his Table 66) consists of affiliative forms of sensory nouns grammaticalized as evidential heads.

The four sensory-evidential formal nouns form a paradigmatic sub-set. Each is the affiliative form of a sensory noun and heads one evidential construction (Nakagawa (2024: 258) Table 66):

The possessed morphology follows the standard affiliative pattern for monosyllabic open-syllable nouns: ru → ruwe(he), haw → hawe(he), and so on Nakagawa (2024: 226).

35.3 Shared diagnostics

Four properties distinguish formal nouns from the other noun subclasses as a group Nakagawa (2022: 480–482).

Obligatory modification. A formal noun cannot constitute an NP on its own. *pe, *hi, and *kur are not grammatical arguments in isolation. This places formal nouns at the foot of the independence hierarchy set out in Nakagawa's Table 48 — below the locative noun's basic form, which at least requires only an immediately preceding NP.

No self-relativization. A formal noun cannot serve as the head of a separate modifying clause in the ordinary sense: *pirka ruwe 'a good track (as NP)' and *pon kur 'a small person (as NP)' are ungrammatical as true relativizations, where an external clause predicates something of the formal noun itself Nakagawa (2022: 480).

No free attributive modifier. When a prenominal demonstrative precedes a formal noun, the result is not a compositional NP but a fused pronoun-like form: tan + pe → tanpe 'this thing', toan + pe → toanpe 'that thing' Nakagawa (2024: 287).

Phonological fusion. p ~ pe and hi fuse prosodically with the preceding element, behaving synchronically like suffixes, even though each nominalizes the entire preceding VP rather than only the adjacent word Nakagawa (2022: 481).

Distribution asymmetry within the class. The antecedent constraint is not uniform: hi requires a VP antecedent; p ~ pe takes a VP or a prenominal but not a bare NP; utar exceptionally accepts a bare NP as well (cikappo utar 'small birds') Nakagawa (2024: 258); Nakagawa (2022: 482).

35.4 The p ~ pe nominalizer

The nominalizer p ‘thing, one’ ~ pe ‘thing, one’ is the most productive member of the set. It attaches to any VP and yields either a relative-clause head (entity reading, with a missing argument) or a content complement (propositional reading, full argument structure) Satō (2008: 174–178); Nakagawa (2024: 503–506). The allomorph is phonologically conditioned: p follows a vowel-final VP; pe follows a consonant-final VP Satō (2008: 177).

As a relative-clause head, p ~ pe nominalizes a VP with a missing argument:

A productive instrumental pattern uses p ~ pe to nominalize an applicative VP: ani X p 'the thing with which one does X', as in ani e=kanpinuye p 'pencil' (literally 'the thing with which you write') and ani humi a=aste p 'musical instrument' (nukopoli, aynu-corpora Discord, 2026-05-29) aynu-corpora Discord (2023–2026). The instrumental ani fills the applicative argument slot while p nominalizes the whole VP ‹corpus-confirmed›.

When a prenominal demonstrative precedes pe, the result is a demonstrative pronoun: tanpe 'this (thing)', neap 'that previously mentioned thing', and the rest of the series treated in Chapter 50 (Demonstratives, Anaphora, and Definiteness) Nakagawa (2024: 287). The full syntax of p ~ pe — its arity-based relativization/complementation split and clause-linker interactions — is in Chapter 101 (The Nominalizers -p/-pe and -i/-hi (Participant, Event, Place, Fact)).

35.5 hi and the other general formal nouns

The formal noun hi ‘thing, matter, time, place’ accepts only a VP antecedent and characteristically nominalizes the state of affairs described by the clause — the content-complement reading — without a missing argument Nakagawa (2022: 482).

hi also supplies the morphological base of several personal pronouns. The first-person singular kani derives from ku=an-hi 'the thing/fact of my existing', reduced by regular vowel contraction; c(i)oka(y) 'we (exclusive)' from ci=oka(y) + reduction; and asinuma from a= + sinuma Nakagawa (2024: 253); Kindaichi & Chiri (1936). The full pronoun paradigm is in Chapter 49 (Personal, Reflexive, and Emphatic Pronouns).

hike 'the one (that …)' is an anaphoric relative head derived from hi by the concretizing suffix -ke — the same element that yields arke 'half' and sake 'refrain-melody' from their bases Nakagawa (2024: 247–248). The suffix -ike that appears in lexicalized items such as onneike 'parent/elder' (< onne 'old' + -ike 'person/entity', per Chiri's gloss; noted by antitwilight, aynu-corpora Discord, 2024-11-15) aynu-corpora Discord (2023–2026) may represent the same element in a lexicalized compound; the synchronic productivity of bare -ike as a nominalizing suffix is a proposed analysis ‹contested›.

katu 'state, manner' nominalizes the manner or condition described by a clause; it appears in the frame … katu ene an hi 'such is the state of affairs' as a manner complementizer Nakagawa (2024: 504). uske 'place' and usi(ke) 'time/place' head NPs denoting the location or occasion of an activity:

The Tokachi/Ishikari form ihi is a regionally restricted variant confined in attested data to the fixed frame ene an ihi 'such a thing' Nakagawa (2024: 257).

Nakagawa's diachronic hypothesis — that hi is the historical source of the affiliative suffix -i (as in sik-i(-hi) 'its eye') — is discussed under Chapter 45 (Morphophonology of the Affiliative Suffix and Echo/Copy Vowels) ‹speculative› Nakagawa (2022: 493).

35.6 kur and kuni: person and purpose nominals

The formal noun kur ‘person’ heads NPs denoting a person or agent. As a relativizer it takes a VP antecedent. As a class-term head it combines with the adnominal un to form demonyms: Sar un kur 'a Saru person', Iyoci un kur 'a Yoichi person' Nakagawa (2024: 234). In the Hokkaido corpus kur ranks at position 98 (2,461 tokens), consistent with its dual role as agent-denoting nominalizer in narrative and as a class-term head.

The formal noun kuni ‘(in order to …; that …)’ is the purpose and intention complementizer. It follows a finite clause and establishes its content as a goal or the object of a mental-state verb: with the transitive verb ramu 'think/believe', the correct embedding is kuni a=ramu 'think that …', contrasting with the quotative frame sekor a=ye 'say that …', which selects a directly reported assertion rather than a propositional attitude (nukopoli, aynu-corpora Discord, 2024-11-09) aynu-corpora Discord (2023–2026) ‹corpus-confirmed›. The related kunak ‘(lest …; whether …)’ carries a negative-purpose or indirect-question reading. Satō (2008: 175) includes both in the nominalizer list alongside hi and p ~ pe. Their full syntax — interaction with purpose adverbials, desiderative and abilitative predicates, and control — is in Chapter 107 (The kuni Complementizer and Purpose Complements).

35.7 The sensory-evidential quartet

ruwe, siri, hawe, and humi are the affiliative forms of four sensory nouns and constitute the formal-noun stratum of the evidential system Nakagawa (2024: 258); Tamura (1996: s.v. ruwe, siri). Each requires a finite VP antecedent, none stands alone as a predicate argument, and each is followed by a predicate — typically the copula ne — that closes the evidential construction.

A morphological point distinguishes evidential use from the full lexical affiliative: the lexical form may carry an additional extension (ruwehe, humihi), whereas the evidential formal noun blocks it — *ruwehe ne is ungrammatical as the inferential evidential Tamura (1996: s.v. ruwe).

Dialect variation affects the morphological shape: southern Hokkaido (Saru, Chitose) uses the possessed forms ruwe / siri / hawe / humi, while Tokachi and Ishikari often use the bare concept forms ru / sir / haw / hum in the same evidential construction Nakagawa (2024: 390). Both shapes head the same frame.

The semantic distinctions among the four — the source-type hierarchy, the indirect evidential frames (siri an, hawe as), and the scope of the evidential over the embedded TAM system — are treated in the dedicated chapters: 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). The architecture common to all four is in Chapter 120 (The Nominalization-plus-Copula Evidential Schema).

35.8 Grammaticalization: from lexical noun to formal noun

The sensory-evidential set makes the grammaticalization cline transparent. The lexical noun ru ‘track, path’ denotes a physical trace; its affiliative form ruwe 'the trace of —' extends to any trace including an epistemic one; clause-final [clause] ruwe ne then asserts the clause content as established fact Tamura (1996: s.v. ruwe); Nakagawa (2024: 258, 392–396). The parallel paths for sir, haw, and hum are transparent from the inventory table. The result is only partial grammaticalization: the lexical source remains recoverable and the choice of post-nominal predicate (ne vs an) still matters, placing these forms at an intermediate point on the cline Nakagawa (2024: 393).

The non-evidential formal nouns stand at a later stage. hi has no independent lexical noun use in modern Hokkaido Ainu; its original sense survives only in the pronoun-formation structures described above and, by Nakagawa's hypothesis, in the affiliative suffix -i. p ~ pe is likewise fully bound; its affiliative-form origin in the monosyllabic open-syllable class (cf. ru → ruwe(he), ci → ciye(he)) is preserved in the paradigm of monosyllabic nouns but no longer lexically active Nakagawa (2024: 226).

The two-stage path — lexical noun → possessed formal noun → complementizer or evidential head — is the mechanism through which Hokkaido Ainu builds clause-embedding without infinitives or dedicated particle complementizers. The point has been noted in the community discussion (nukopoli, aynu-corpora Discord, 2024-11-25) aynu-corpora Discord (2023–2026): in a head-marking language where every embedded clause carries its own person marking, the formal noun provides the NP wrapper that the embedding construction requires, substituting for the infinitive strategies of other language types ‹contested›. The full treatment of relativization, complementation, and the arity diagnostic is in Chapter 101 (The Nominalizers -p/-pe and -i/-hi (Participant, Event, Place, Fact)), Chapter 106 (Nominalized Complements and Control/Raising), and Chapter 102 (Lexical-Head and Formal-Noun Nominalization (kur, uske, ruwe/hawe/siri/humi)).