Chapter 120The Nominalization-plus-Copula Evidential Schema

120.1 The nominalization-plus-copula schema

Hokkaido Ainu encodes the speaker's source of information through a productive construction: a finite verb phrase is nominalized by a bound formal noun and the result is predicated with the copula ne or an existential or posture verb. The language has no dedicated evidential verbal affix. The four formal nouns that build the construction — ruwe, siri, hawe, and humi — are each the possessed form of a common sensory noun, and they carry their sensory origin into their evidential meaning Nakagawa (2024: 258, 390–392); Hirosawa (2026: 64).

The structural template is [CLAUSE]-FORMAL.NOUN + ne: the nominalized event is presented as the possessed trace, appearance, voice, or feeling of something, and the copula closes the predication. The formal noun specifies the information channel; the copula anchors the proposition to the speaker's direct or reflective perspective. Four direct evidential forms use the copula ne; a parallel set of indirect counterparts replaces it with the existential an, the action verb ki, or the posture verb as Dal Corso (2018: 304–314). The construction belongs to the broader nominalization system treated 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 formal-noun word class that makes it possible is defined in Chapter 35 (The Formal / Defective Noun Set (ruwe, hawe, siri, humi, hi, pe, kur)).

120.2 The four formal nouns: morphological sources and evidential values

Tamura's Saru dictionary records the source nouns, their regular possessed forms, and the evidential formal nouns derived from them Tamura (1996: s.v. ru, sir, haw, hum, ruwe, siri, hawe, humi). Nakagawa tabulates the evidential values Nakagawa (2024: 258).

The regular possessed forms of the source nouns may carry an additional ending — ruwe(he), humi(hi) — but the evidential formal nouns block this extension: *ruwehe and *humihi are ungrammatical in the evidential function Tamura (1996: s.v. ruwe, humi). Whether the concept form (sir, haw) and the corresponding possessed form carry a usage or nuance difference is explicitly left unresolved in Tamura's dictionary Tamura (1996: s.v. sir) ‹speculative›. The possessed morphology is significant in its own right: a noun in its possessed form grammatically presupposes an antecedent, so these nominalizers formally encode that the speaker already has the relevant situation in hand Satō (2008: 135).

120.3 Typological placement: the C1 system

Mapped onto Aikhenvald's evidential typology Aikhenvald (2004: 63–65), the Ainu system belongs to the C1 type: visual, non-visual sensory, inferred, and reported sources of information are grammatically distinguished Hirosawa (2026: 64); Takahashi (2013: 129). The type is realized analytically through a nominalization-plus-copula construction, not through dedicated verbal affixes — this makes Ainu a textbook case of the nominalization-based evidential strategy Dal Corso (2018: 24–25); Takahashi (2022) ‹consensus›.

The cell assignments are approximate. ruwe ne covers the inferred cell (information from reasoning or assimilated knowledge), siri ne the visual cell, humi ne the non-visual sensory cell, and hawe ne the reported cell. The boundary between the visual (siri) and the knowledge-from-trace (ruwe) cells is contested in the recent literature and is discussed in the ruwe ne chapter Dal Corso (2018: 311–312); Hirosawa (2026: 81) ‹contested›.

120.4 Direct and indirect evidentiality: the eight-form system

Dal Corso's morphosyntactic account resolves the four formal nouns into an eight-form system organized by two parameters: source reliability (REL) and event accessibility (ACCES) Dal Corso (2018: 304–337). The four direct forms use the copula ne (internal, speaker-anchored perspective); the four indirect forms replace it with the existential an, the action verb ki, or the posture verb as (external perspective; 2018: 313–314). Source reliability is not epistemic modality: it specifies the psychophysical or cognitive channel of acquisition and says nothing about the speaker's degree of certainty Dal Corso (2018: 339).

The underlying sense hierarchy runs Sight > Smell/Taste–Touch > Hearing, yielding a coarser visual/non-visual split: ruwe ne and siri ne are the higher-reliability direct members, humi ne and hawe ne the lower Dal Corso (2018: 312). A structural asymmetry distinguishes the ru-series: in Saru, Chitose, and Biratori the trace-noun enters only direct evidentiality, and there is no indirect *ru an counterpart to siri an Dal Corso (2018: 309, 326). The hawe as indirect form has, moreover, nearly fully specialized as a reportative — for quotation when the source is overt and for hearsay when it is absent — while humi as is never reportative Dal Corso (2018: 328–330). The per-form indirect semantics are treated in 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).

120.5 Token frequencies

ruwe ne dominates the paradigm quantitatively. In Dal Corso's Hokkaido reference texts it accounts for 1508 of 2232 evidential tokens — roughly two thirds of all eight forms, and more than five times the next most frequent form (hawe ne, 298 tokens) Dal Corso (2018: 72). The asymmetry is independently noted by Tamura and by Yoshikawa: ruwe ne outstrips the other forms in oral-narrative (口承文芸) registers, where it functions as the default frame-text predicate Hirosawa (2026: 73). Community discussion characterizes ruwe ne in the same terms — frequent especially in oral narratives and legends, functioning like Japanese のだ to close narrative segments and mark topic shifts (nukopoli, aynu-corpora Discord, 2024-11-09) aynu-corpora Discord (2023–2026).

These counts derive from Dal Corso's specific reference texts (Kayano, Tamura, Nakagawa, Bugaeva), not from the full Hokkaido corpus, and should be treated as indicative of the skew rather than as absolute proportions Dal Corso (2018: 41) ‹corpus-suggested›. The dominance of ruwe ne is consistent with its status as the unmarked, most grammaticalized member of the paradigm. The genre distribution of evidential usage across narrative registers is treated in Chapter 155 (Narrative TAM and Evidential Patterning by Genre).

120.6 Analytical frameworks

Four analytical traditions describe the same construction from different angles, and they partially conflict.

The oldest is the mood (法) tradition. Kindaichi identified seven distinctive moods in Ainu; within that taxonomy siri is the 見説法 (seeing-speech mood), hawe the 聞説法, humi the 感説法, and ruwe the 確説法 / 事実法 Kindaichi & Chiri (1936); Hirosawa (2026: 65–68). The mood label captures the speaker-oriented, assertive character of these forms but does not isolate their source semantics.

The modern evidentiality framing, now standard in the descriptive literature, places the system in the typological framework of information-source encoding (Aikhenvald 2004). It is adopted by Takahashi Takahashi (2013: 129), Dal Corso Dal Corso (2018), Nakagawa Nakagawa (2024: 392), and Hirosawa Hirosawa (2026: 64). Dal Corso's is the most detailed morphosyntactic account (§Direct and indirect evidentiality above).

The predication-type framing (Takahashi 2013) adds a dimension orthogonal to source channel: in Tokachi, the visual nominalizer sir is restricted to stage-level (transient, scene-describing) predication, and individual-level (permanent) properties take ru regardless of the information channel Takahashi (2013: 131–133). The ontological type of the predicate thus independently constrains the available evidential ‹consensus›.

The information-structure framing is Hirosawa's (2026). He argues that ruwe ne additionally marks discourse-new information, parallel to Japanese のだ, and unconditionally declares the content to lie inside the speaker's territory of information Hirosawa (2026: 75–81, 89). Dal Corso explicitly argues that the territory-of-information account is inapplicable to Hokkaido direct evidentiality Dal Corso (2018: 335–337). These two recent treatments disagree on the right framework for ruwe ne ‹contested›. Community discussion, summarizing Yoshikawa (2020), characterizes ruwe ne as semantically unmarked relative to the other forms — lacking their specific evidentiary highlighting — and as functioning to close a narrative segment and transition to the next (場面転換・統括), consistent with Hirosawa's analysis (nukopoli, aynu-corpora Discord, 2024-03-17) aynu-corpora Discord (2023–2026); this observation is treated here as corroboration of the literature rather than an independent authority.

Ijäs's pedagogical grammar presents the system under the evidentiality framing without adopting any of the specific theoretical positions above Ijäs (2023). The pedagogical description aligns with the standard evidential account and is useful for language learners but does not adjudicate the analytical disputes.

120.7 The shared clause-final frame paradigm

All four formal nouns participate in a shared set of clause-final frames that vary the following element and the sentence type. Nakagawa arranges these into a paradigm of four patterns Nakagawa (2024: 392–396):

The bare polar-question frame is an instance of insubordination — a once-embedded nominalized clause promoted to main-clause use Takahashi (2013: 132). Negation operates inside the evidential scope: the predicate is negated with somo before the formal noun, while the formal noun remains outside the scope of negation (…somo ki ruwe ne 'it is a fact that [one does not …]') Nakagawa (2024: 393). The interrogative frames are treated in Chapter 144 (Interrogative Strategies, Question Particles, and Evidential Questions); the existential frame in the existential an chapter.

120.8 Scope over TAM and the finiteness of the embedded clause

The constituent nominalized by the formal noun is a finite, clause-sized predication. It carries its own person marking, verb suppletion (singular ek vs. plural arki 'come'), aspectual auxiliaries, perfective a, and negation; the evidential formal noun scopes over that whole TAM-marked clause Nakagawa (2024: 393). The layering is transparent from ordering: the formal noun may be followed by aspectual material (a, nankor), final particles (na, wa), and clause-linkers (hine, kusu), but the intentive kusu ne ('intends to') cannot follow the evidential: kusu ne ruwe ne is well-formed but *ruwe ne kusu ne is not Nakagawa (2024: 393). The evidential thus sits above clause-internal aspect but below clause-linkers and intentive modality. Community discussion independently observes that this ordering follows from the clause hierarchy placing the evidential inside modality operators (nukopoli, aynu-corpora Discord, 2024-11-09) aynu-corpora Discord (2023–2026).

Dal Corso ties the copula vs. existential/posture-verb choice to perspective: the copula ne expresses an internal, speaker-anchored viewpoint and hosts the direct forms; an, as, and ki express an external viewpoint and host the indirect forms Dal Corso (2018: 313–315). A structural polysemy follows for siri an: it is three-way ambiguous between the interrogative existential frame, the indirect visual evidential, and an aspectual resultative-progressive — a polysemy treated in Chapter 122 (siri ne — the Direct-Perception / Situational Evidential) and Chapter 112 (The 'Appearance' Continuative: siran, siri…, and kane an). The interaction of the evidential with the wider clause-final TAM and negation apparatus is treated in Chapter 100 (TAM, Evidentiality, and Negation on Nonverbal Predicates) and Chapter 96 (The Copula ne and Predicate-Nominal Clauses).

120.9 Predication type and the Tokachi stage-level constraint

In Tokachi the choice among the formal nouns is constrained not only by information source but also by the ontological type of the predicate. Takahashi's Tokachi study shows that the visual nominalizer sir is restricted to stage-level (transient, scene-describing) predication; for individual-level (permanent property) predicates — even those known by sight — Tokachi uses ru, which is unmarked for information type Takahashi (2013: 131–133). Of the four, only sir carries this restriction; hum, haw, and ru function at both predication levels.

Takahashi parallels the constraint with the ban on individual-level complements of English perception verbs; in Ainu the restriction surfaces on the nominalizer rather than on a matrix verb. This finding sets bounds on any pure source-channel account of the formal-noun choice: the grammar of predication independently constrains the available evidential ‹consensus›. Its implications for the ruwe ne / siri ne boundary are discussed in the ruwe ne chapter.

120.10 Grammaticalization and the nominalization-to-evidential cline

The formal nouns are partially, not fully, grammaticalized from their source nouns. The diachronic path is already stated in the early grammars: the source nouns function first as lexical nouns, then as clause-nominalizing heads in embedded complements, then as clause-final evidential sentence-enders Takahashi (2013: 134); Takahashi (2018: 111). At each stage the lexical semantics of the source noun remains visible, but the syntactic freedoms of a full noun are progressively lost. The evidential formal nouns admit no demonstrative or attributive modifier of their own, cannot head an ordinary common-noun relative clause, and are fixed in linear position Dal Corso (2018: 24–25); Nakagawa (2024: 393).

Dal Corso stresses that the degree of grammaticalization varies across the system: ruwe ne is the most grammaticalized of the four Hokkaido direct forms, with the Sakhalin particle manu at the far end of the cline as a fully fused suffix Dal Corso (2018: 24–25). The Sakhalin nominalization innovations and their contrast with Hokkaido are set out in (Bugaeva 2016). The path from noun to nominalizer to clause-final particle belongs to the wider nominalization-grammaticalization cline treated in Chapter 102 (Lexical-Head and Formal-Noun Nominalization (kur, uske, ruwe/hawe/siri/humi)).

120.11 Dialect microvariation

The semantic organization of the system — four channels, direct vs. indirect — is consistent across Hokkaido; its morphosyntactic realization varies. Four dialect-level patterns stand out.

• Saru, Chitose, Biratori (the core of Dal Corso's data) use the possessed forms ruwe / siri / hawe / humi and maintain the eight-form direct/indirect system Dal Corso (2018: 304–337).
• Tokachi (and frequently Ishikari) close the clause with the bare concept forms ru / sir / haw / hum rather than the possessed forms Nakagawa (2024: 110, 390); Takahashi (2013: 129–131). Allomorphs sin (for sir) and aw (for haw) are attested Takahashi (2013: 130–131).
• Horobetsu grammaticalizes the possessed-form nominalizers a step further into adverbial-clause markers and conjunction-like particles — a use not reported for well-described Saru Takahashi (2018: 107–112). Kindaichi and Chiri had already noted the nominalizers in conjunction-like uses; Takahashi shows Horobetsu has systematized this extension, bringing evidentiality into the set of categories expressed in adverbial clauses alongside aspect and modality — an instance of 'versatile nominalization' Takahashi (2018: 112) ‹corpus-suggested›.
• Asahikawa and Shiranuka show an inferential ru an, with the trace-noun entering the indirect frame in apparent complementary distribution with the siri an of the southern dialects Dal Corso (2018: 326). This is the one Hokkaido context where ru forms indirect evidentiality; the distribution is flagged by Dal Corso as requiring a wider-dialect study ‹corpus-suggested›.

Dialect coverage is solid for Saru/Chitose/Biratori and Tokachi; thin or second-hand for Ishikari, Asahikawa, Shiranuka, and Shizunai. The dialect classification is set out in the dialectology chapters; the TAM-evidential variation across registers is in Chapter 155 (Narrative TAM and Evidential Patterning by Genre).

120.12 Double evidentiality

Two evidential expressions may stack in one clause; this is more frequent in Hokkaido than in Sakhalin Dal Corso (2018: 331–333). Two constructions recur. In the first, a character's indirect source is placed within the narrator's ruwe ne, yielding both the character's subjective source and the narrator's assertion of the event as traditional knowledge:

In the second pattern, ruwe ne reinforces a lower-reliability direct form, boosting epistemic commitment: attested hawe ne nankor ruwe ne 'maybe my god secretly let me know' Dal Corso (2018: 332). Other attested combinations include hawe as ruwe ne, siri ki siri ne, siri an ruwe ne, and humi ne humi as Dal Corso (2018: 331–333). The consistent outer position of ruwe ne in stacked constructions supports its analysis as the epistemic-incorporation operator of the paradigm, a role discussed in Chapter 121 (ruwe ne — the Inferential / Visual-Trace Evidential).

120.13 Mirativity and the admirative

The four formal nouns do not themselves encode mirativity (意外性). Dal Corso finds no evidence that the Hokkaido evidential paradigm systematically generates mirative readings; this contrasts with Sakhalin, where personal-knowledge evidentials can trigger mirative interpretations through a deictic-gap mechanism Dal Corso (2018: 336). In Hokkaido, mirative value attaches instead to the perfective auxiliary aan and to anan Hirosawa (2026: 90). Yoshikawa argues that aan signals indirect evidentiality plus mirativity (reported via Hirosawa (2026: 90)) Yoshikawa (2018) ‹speculative› (Yoshikawa 2018 cited second-hand).

The two mechanisms can co-occur. When aan places information outside the speaker's expected knowledge and ruwe ne follows, the sequence expresses surprise then reclamation: the speaker marks the event as unexpected (aan) and then declares it recognized fact (ruwe ne). Hirosawa parallels this with Japanese みたいな + のだ Hirosawa (2026: 90). The full treatment of the evidential-to-mirative interface, evidential scope, and the cline from formal-noun evidentials to sentence-final particles is in Chapter 125 (Mirativity, Evidential Scope, and Grammaticalization to Sentence-Final Particles).