Chapter 119Epistemic Modality, Dubitative, Intentive, and Irrealis

119.1 The epistemic and irrealis modal field

Hokkaido Ainu expresses epistemic modality — the speaker's calibration of the probability or certainty of a proposition — and irrealis modality — marking that an event has not yet been actualized — through periphrastic auxiliaries that follow the lexical verb in the auxiliary frame. In this frame the lexical verb carries person agreement while the auxiliary remains morphologically unmarked, the same pattern as the perfective a and the admirative aan treated in Chapter 111 (The Perfective/Anterior Particle a (a, a…a, aan)). Eight forms occupy this domain, ranging from epistemic probability to volitional commitment to reportative hedging Nakagawa (2024: 306–319); Bugaeva (2012: 493–494). Across the typological literature on grammaticalization, the path from lexical verb to modal auxiliary is well documented, and in Ainu most of these forms retain visible lexical identities in the auxiliary set (Bybee et al. 1994).

The TAM architecture that houses these auxiliaries — the distinction between the auxiliary frame, the pair-verb frame, and the transitive-verb-auxiliary frame — is mapped in Chapter 108 (Tense, Aspect, Mood, and Modality: Overview). The four strategies Ainu employs to recover temporal reference without grammatical tense, among which the prospective auxiliaries play a central role, are treated in Chapter 109 (Tenselessness and the Time-Reference System). Deontic obligation and permission (easirki, yak pirka, yakka pirka) belong to Chapter 118 (Deontic Necessity and Obligation (kuni ne, kus ne)).

119.2 nankor — epistemic probability and conjecture

The auxiliary nankor ‘probably’ follows the person-marked lexical verb and indicates that the speaker regards the proposition as probable, likely, or to-be-expected without asserting it as established fact: 推量「〜だろう」 Nakagawa (2024: 311). Bugaeva characterizes it as "conjective probabilistic mood" Bugaeva (2012: 493). With a future or present-state predicate the reading is probability ('probably will', 'is probably'); with a past-referenced predicate it yields epistemic necessity ('must have'). In the corpus the form returns 1,977 tokens at rank 108 — the highest frequency among the modal auxiliaries — and it appears in narrative, spoken monologue, and chained clauses alike.

With a second-person subject and the sentence-final particle na, nankor shifts from probability to a softened expectation or indirect command — 「〜だろうよ」 — functioning as a milder alternative to the bare imperative Nakagawa (2024: 311, 423); Ijäs (2023). The same reading surfaces in prohibitive contexts: iteki … somo eci=ki nankor na 'you had better not not do it' Nakagawa (2024: 425).

Phonologically, the string nankor na frequently assimilates to nankonna in running speech Nakagawa (2024: 311). The auxiliary can additionally appear after ruwe ne in the post-evidential slot, where it hedges an already-evidentialized proposition as merely probable; this post-nominal position — shared with the perfective a and final particles — is part of the frame inventory for ruwe ne detailed in the clause-final frames section Nakagawa (2024: 393).

Ijäs (Lesson 10) proposes that the modal nankor derives from a lexical verb nankor 'be trustworthy', but marks this derivation as probable rather than established Ijäs (2023) ‹speculative›. No grammar consulted for this chapter traces the auxiliary to that lexical source, and the connection remains unverified.

119.3 kus(u) ne — prospective intentional

The prospective construction consists of the purposive conjunction kusu ‘in order to, so that’ — phonologically reduced to kus in running speech — followed by the copula ne. Person agreement falls on the lexical verb and the kusu ne string is morphologically frozen: the copula is not separately inflected Nakagawa (2024: 435); Satō (2008: 85–86). As one of the four strategies for recovering future time reference without grammatical tense (→ Chapter 109 (Tenselessness and the Time-Reference System)), it marks an event as unrealized at the moment of speaking. Three situational readings emerge from person and discourse context:

• First-person intention: 'I intend to / I plan to' — the speaker commits to an unrealized future action.
• Scheduled future: 'it is to happen / it will happen' — the event is foreseen as certain without necessarily implying personal volition.
• Second-person near-command: with na → 'you should / you're going to' — a directive softer than the bare imperative (→ Chapter 114 (Imperative and Prohibitive (Directive Mood))).

Nakagawa documents all three at 2024: 435, §14.1.2 (exx.966–969); Satō at 2008: 85–86.

Ijäs writes the sequence as a single orthographic word, kusune, and tags it AUX or PRT-FIN; the analysis followed here retains the two-morpheme parse — conjunction kusu plus copula ne — treating the one-word spelling as a pedagogical convention rather than a morphological claim Nakagawa (2024: 435); Ijäs (2023). The element kusu functions independently as a reason or purpose clause conjunction (→ Chapter 135 (Conditional, Temporal, and Concessive Clauses)); the prospective auxiliary is the specific combination kusu + ne, with the copula predication turning an unrealized purpose into a future-directed commitment.

The scope of kusu ne is strictly clause-internal: it sits below the evidential frame, so kusu ne ruwe ne is well-formed ('it is a fact that one intends to') while *ruwe ne kusu ne is ungrammatical Nakagawa (2024: 393). The interaction with the evidential system is developed in the scope section below.

119.4 Near-future auxiliaries: oasi, etokus, etokoyki

Three additional auxiliaries extend the prospective periphery beyond kusu ne Nakagawa (2024: 316–318, 435):

• oasi ‘will certainly happen’ — 'surely will happen' (将来確実に〜が起こる); a stronger certainty than the merely probable nankor. With tane 'now' it yields an imminent reading, 'is just about to' Nakagawa (2024: 318).
• etokus ‘be about to’ — 'be about to (do)': the event is on the point of beginning Nakagawa (2024: 316).
• etokoyki ‘be getting ready to’ — 'be getting ready to': the agent is actively in the preparatory stage for the action Nakagawa (2024: 435).

All three are catalogued in Nakagawa's 表74 (本来的な助動詞, the primary Saru-dialect auxiliary inventory) and in Bugaeva's English inventory Nakagawa (2024: 306, 316–318, 435); Bugaeva (2012: 493–494). No banked corpus examples are currently available for these three forms beyond the grammar descriptions; the claim that they carry the stated meanings rests on Nakagawa's documented paradigm ‹corpus-suggested›.

119.5 Desiderative rusuy and expectation kuni

rusuy ‘want to’ is the desiderative auxiliary: 'want to do'. The lexical verb takes the person affix and rusuy follows unmarked; a strong tendency toward phonological incorporation produces forms such as arparusuy 'want to go' and iperusuy 'want to eat' Nakagawa (2024: 312–313); Bugaeva (2012: 493). Argument structure, distribution with negation, and the incorporated variants are treated fully in Chapter 116 (The Desiderative rusuy).

A typological comparison has been proposed in the community: nukopoli (aynu-corpora Discord, 2024-11-29) observes that rusuy parallels Mandarin Chinese 要 (yào), which similarly covers both desiderative ('want to') and near-future readings, while kusu ne resembles English will in combining intentionality with temporal commitment aynu-corpora Discord (2023–2026) ‹speculative›. No grammar read for this chapter develops this typological comparison.

kuni ‘should, be supposed to’ straddles the epistemic and deontic domains. As an auxiliary it marks expectation or obligation ('should, be supposed to, it is expected that'); as a subordinating particle it introduces complement clauses; and as a formal noun it heads nominalized purpose structures — a=an kuni uske 'a place where one should be' (Nakagawa 2024: 310, ex.461). The deontic uses and the nominalized-complement function are treated in Chapter 118 (Deontic Necessity and Obligation (kuni ne, kus ne)) and Chapter 106 (Nominalized Complements and Control/Raising) respectively.

119.6 manu — reportative hedging

manu ‘reportedly, so-called’ is listed among the primary auxiliaries (本来的な助動詞) at Nakagawa 2024: 310–311 (exx.464–467). Its core value is attributive or reportative hedging: 'the so-called, reportedly, it is said that' — the speaker presents the proposition as attributed, named, or received rather than directly asserted.

Nakagawa observes that manu approximates the Sakhalin sentence-final hearsay marker — the cognate form in Sakhalin Ainu — while remaining a distinct Hokkaido auxiliary, not a particle Nakagawa (2024: 311). Within the Hokkaido system it sits on the boundary between the epistemic modals treated here and the reportative evidential centered on hawe ne (→ Chapter 123 (hawe ne — the Reportative / Hearsay Evidential)). Unlike hawe ne, which nominalizes a full clause and predicates it as verbal report, manu inserts inside the predication and signals attribution of a name or inherited category rather than a cited information channel. Corpus token counts are currently unreliable for manu given surface-form homophony; it is treated as infrequent in the Saru record.

119.7 Scope and the evidential interface

The epistemic and irrealis auxiliaries treated in this chapter all occupy clause-internal scope — below the evidential frame. The ordering constraint is strict: kusu ne ruwe ne is well-formed (the prospective claim is nominalized and predicated as established fact), while *ruwe ne kusu ne is ungrammatical because the prospective auxiliary cannot appear in the post-evidential position Nakagawa (2024: 393). Negation belongs inside the frame in the same way: somo ki ruwe ne negates the verb before the formal noun, not after it. The full account of this layering appears in the scope and finiteness section.

Members of the aynu-corpora community have independently noted the same asymmetry: nukopoli (aynu-corpora Discord, 2024-11-09) observes that while ruwe ne kusu (reason clause: 'because it is a fact that…') remains well-formed, *ruwe ne kusu ne as a prospective construction after an evidential is ruled out, attributing this to a clause hierarchy in which modality must remain inside the outer scope of the evidential aynu-corpora Discord (2023–2026); Nakagawa (2024: 393).

Corpus data supply chains that make the scope layering concrete. The most common multi-auxiliary configuration places the prospective inside the admirative aan, which is in turn nominalized by ruwe ne — a three-level stack in the order prospective ≺ mirative ≺ evidential.

nankor behaves differently from kusu ne in this respect. As an epistemic auxiliary it can appear either before ruwe ne — inside the evidentialized proposition ('it is a fact that X probably…') — or after it, in the post-evidential slot where it hedges the evidentialized claim as merely probable rather than fully asserted Nakagawa (2024: 393). This dual positional freedom is shared by the perfective a and final particles, both of which may follow ruwe ne without being embedded inside the proposition; the full inventory of post-nominal elements is set out in the clause-final frames section.

The conjunction kusu in conditional, purpose, and reason clauses — as distinct from the prospective kusu ne — is treated in Chapter 135 (Conditional, Temporal, and Concessive Clauses). The full interface between the modal auxiliaries and the evidential paradigm, including stacking patterns and the evidential-to-mirative cline, belongs to Chapter 125 (Mirativity, Evidential Scope, and Grammaticalization to Sentence-Final Particles).