Chapter 150First-Person Narration, the Logophoric Fourth Person, and Reported Discourse

150.1 The reported-discourse system in overview

Hokkaido Ainu marks reported discourse with a quotative index (QI): a quotative marker that follows the quoted string and is in turn followed by a speech or cognition verb. The central marker is sekor, an adverbial conjunction Bugaeva (2008: 35–36) that is obligatory in direct and semi-direct constructions. A structurally separate class of constructions uses bound complementizers (kunak, yak, kuni) with transitive reporting verbs and is treated as a complementation device in Chapter 105 (Quotative Complementation and Reported Speech (sekor)) and Chapter 107 (The kuni Complementizer and Purpose Complements). This chapter covers the sekor-mediated system: the transitivity split among reporting verbs, the direct/semi-direct/indirect distinction, the 1st-to-4th-person shift inside quotes, the logophoricity debate, the first-person narration convention of yukar and uwepeker, and the hawe family of forms that bridges the quotative and evidential systems. The evidential architecture within which hawe ne and hawe as sit is described in Chapter 120 (The Nominalization-plus-Copula Evidential Schema) and Chapter 123 (hawe ne — the Reportative / Hearsay Evidential).

The quotative and evidential systems intersect at the noun haw ‘voice’: the intransitive report verb hawean ('a voice exists' → 'say'), the reportative evidential hawe ne ('it is a voice/report that'), and the hearsay form hawe as ('a voice arises') all derive from the possessed form haw-e 'the voice of —' Nakagawa (2024: 398, 428). Reference-tracking across narrative spans, which the quotative machinery helps manage, is described in Chapter 139 (Reference Tracking and Argument Continuity).

150.2 sekor and the reporting-verb split

The etymology of sekor is transparent: it derives from se 'make a sound / say' + kor 'while / while holding', yielding roughly 'while saying' Nakagawa (2024: 470) (nukopoli, aynu-corpora Discord (2023–2026), 2024) ‹corpus-confirmed›. In the southern Hokkaido dialects it is fully grammaticalized as the quotative particle and no longer carries its compositional meaning. The form following sekor is the reporting verb, and its transitivity determines the construction type.

The constraint was first stated by Chiri (1942: 572–573) and systematized by Bugaeva (2008: 36–37): sekor co-occurs with intransitive reporting verbs because the quoted string is adverbial — it is not the object of the reporting verb. The complementizer constructions (kunak/yak/kuni) take transitive verbs (ye 'say', ramu 'think') because the complement clause is the object. The community discussion confirms this: *sekor a=ramu is ungrammatical — ramu requires a nominal object; the correct form is kuni a=ramu or, with the intransitive cognate, sekor yaynu=an (nukopoli, aynu-corpora Discord (2023–2026), 2024) ‹corpus-confirmed›.

The apparent exception is sekor a=ye: the transitive verb ye occurs after sekor in a lexicalized frame that is more frequent in the corpus than the intransitive alternative sekor itak=an (nukopoli, aynu-corpora Discord (2023–2026), 2024; thegodofneet, aynu-corpora Discord (2023–2026), 2025) ‹corpus-confirmed›. Nakagawa (2024: 470) acknowledges sekor a=ye as a lexicalized exception to the transitivity constraint. Ijäs (2023) presents both sekor hawean and sekor a=ye as the two standard quotative frames in her pedagogical grammar.

150.3 Direct, semi-direct, and indirect reported discourse

Bugaeva (2008: 34–35) identifies three construction types for Hokkaido Ainu, distinguished by the treatment of person inside the quote and by the syntactic integration of the complement. The table organizes the system:

The DRD/SDRD distinction rests on person: when the current narrator quotes herself, 1st-person forms inside the quote are retained, and the construction is direct. When she quotes a different speaker, the original 1st-person of the quoted speaker is replaced by the 4th/indefinite person — the pronoun asinuma for the singular logophoric-like reference, the prefix a= on transitive stems, and the suffix =an on intransitive stems — and the result is semi-direct Tamura (2000: 75); Bugaeva (2008: 37–40). The free pronoun asinuma and the 4th-person affix paradigm are described in Chapter 49 (Personal, Reflexive, and Emphatic Pronouns) and Chapter 67 (The Indefinite/Fourth Person: Forms, Reference, and Agent-Defocusing).

Tamura (2000: 75) is explicit that the 4th-person-in-quote is not indirect narration. Bugaeva (2008: 38–43) formalizes a semi-direct category: the syntax of the quote is direct, but the person form signals that the source of the quoted content is a participant other than the current narrator. Nikitina & Bugaeva (2021) sharpen the analysis: across languages, logophoric speech shares its syntax with direct speech and none with indirect; the Ainu semi-direct construction is syntactically direct-like and its person shift is a logophoric device, not a symptom of syntactic subordination ‹contested›. The clause-integration mechanics of the three types are treated in Chapter 131 (Clause Linkage in Ainu: Typology and Inventory).

150.4 The logophoricity debate

Two analyses of the 4th-person-in-quote compete. Tamura (1972) calls it the 'quotative first person' (引用の1人称), treating it as a sub-use of the 4th person alongside the inclusive, impersonal, and honorific uses. Nakagawa (2024: 179) lists the quotative use as use ⑤ of the 4th person without positing a dedicated logophoric category. Bugaeva (2008: 40–43) argues for a genuinely logophoric analysis: the pronoun asinuma (literally a=sinuma 'the self of one') is the singular logophoric form, invariably glossed by the speaker Mrs. Ito Oda as じぶん 'self' rather than わたし 'I'; it tracks the original speaker of the quoted discourse, not the current narrator ‹contested›.

Satō (2008: 259–262) and Satō (2004: 179–184) treat the same forms as polysemous, deriving the narrative-first-person use from a first-person-plural (inclusive) strategy in which the 4th-person affixes index both protagonist and audience. The whole-range of the 4th person and the competing analyses of its referential extent are laid out in Chapter 68 (Honorific and Logophoric Uses of the Fourth Person).

The available evidence leaves several points unsettled. The logophoric analysis is based on Saru–Chitose data (Mrs. Ito Oda, Biratori); whether asinuma carries the same dedicated 'original speaker' meaning in other Hokkaido dialects is unverified ‹speculative›. Nakagawa (2024: 471) notes the same dialect gap. The best-supported claim is: in southern Hokkaido, when a narrator quotes a speaker other than herself, the quoted speaker's 1st-person morphology is systematically replaced by the 4th-person set, and this replacement is not optional ‹corpus-confirmed›. Whether the replacing forms constitute a grammaticalized logophoric subsystem or are a deployment of the general 4th person is the contested analytical choice.

150.5 Yukar and uwepeker: narration in the protagonist's voice

The narrative first-person convention of yukar (heroic and divine epic) and uwepeker (prose tale) consists in the reciter voicing the protagonist through the 4th/indefinite person forms (a=, =an, i=) rather than the narrator's own 1st-person forms Nakagawa (2024: 174–179); Satō (2004: 179–184). The mechanism is structurally parallel to the semi-direct shift in §3: the protagonist's perspective is indexed by the 4th-person set throughout the tale, while any stepping-out narrator commentary, where it appears, uses the 1st person. The oral-literature genres and their registers are described in Chapter 149 (The Oral-Literature Genre System and Its Grammatical Signatures), and the TAM and evidential patterns by genre in Chapter 155 (Narrative TAM and Evidential Patterning by Genre).

The analytical question is whether the convention is a genre-bound pragmatic device or a grammaticalized logophoric system. Several observations favour a convention-first reading: (a) the 4th-person forms are identical to those marking indefinite agents in everyday speech; (b) the distribution skews heavily by genre — the pattern is densest in yukar, lighter in uwepeker, lighter still in everyday quoted speech Nakagawa (2024: 176–177) ‹corpus-suggested›; (c) Satō (2004: 183) derives the narrative use from the inclusive reading, on which the protagonist is voiced as 'we' (protagonist and audience together). Against this, Bugaeva (2008: 42–43) notes occurrences of asinuma as the logophoric form outside oral literature, which would push toward a grammatical analysis that is not genre-defined ‹contested›. The chapter's conclusion is that a genre-convention analysis accounts for the distribution but does not preclude a logophoric analysis of the form: the two claims are compatible if the logophoric device originated in the narrative genre and extended into everyday quotation ‹speculative›.

150.6 The hawe family: report verb, reportative evidential, and hearsay

All three members of the haw ‘voice’ family derive from the possessed form haw-e 'the voice of —'. They differ in the predicate that follows the possessed noun and, correspondingly, in what they assert about the source of information:

The reportative evidential hawe ne differs from the quotative frame in scope: sekor … hawean presents a specific attributed speech event, while hawe ne predicates the whole proposition as something the speaker knows through words, without identifying a specific utterer Nakagawa (2024: 398). The hearsay form hawe as extends further: it reports ambient or indirectly received information, the 'sound' of something being said rather than a specific attributed utterance:

The naming cleft X sekor a=ye hawe ne 'it is what is called X — so it is said' layers both systems. The inner sekor a=ye lexicalized frame embeds the name; the outer hawe ne frames the whole as a conventionally reported fact:

150.7 The sekor a=ye p ne naming cleft

A related but structurally distinct naming frame is X sekor a=ye p ne 'it is the thing that is called X'. Here p is the nominalizer 'thing / the one', and ne is the copula (see Chapter 101 (The Nominalizers -p/-pe and -i/-hi (Participant, Event, Place, Fact))). The community discussion raises a valency point: in this frame the nominalizer p is zero-valent — it is not the object of ye. The transitive verb ye (+2) has its two valency slots filled by the agent a= (−1) and the implicit 3rd-person zero-object (−1), and p then nominalizes the whole clause from outside (nukopoli, aynu-corpora Discord (2023–2026), 2024) ‹contested›. If correct, the frame has the same underlying structure as the general predicate-nominalization cleft described in Chapter 101 (The Nominalizers -p/-pe and -i/-hi (Participant, Event, Place, Fact)), with sekor embedded inside.

150.8 Dialect variation: sekor and ari/ani

Sekor is the quotative marker of the southern Hokkaido dialects. Other Hokkaido varieties use ari or ani in the same quotative slot. Nakagawa (2024: 471–472) notes the dialect distribution. The community discussion proposes that ari/ani is the older pan-Ainu quotative and that sekor is a Saru-region innovation: an analytical form built from se + kor that replaced ari/ani as the Saru dialect grew as the cultural centre (nukopoli, aynu-corpora Discord (2023–2026), 2024) ‹contested›. The directionality of change is not independently datable from the available sources, and the claim rests on internal evidence (the composition of sekor) and the dialect distribution, not comparative-historical reconstruction. The equivalent in Sakhalin varieties (nah), occupying the same functional slot, is cited for typological contrast only; its morphology and distribution belong to the Sakhalin contrast chapter.

150.9 Honest gaps

Three significant uncertainties bear on the analysis of this chapter. First, the logophoric vs. 4th-person debate (§4) remains open: the key evidence for a dedicated logophoric pronoun asinuma comes from Saru–Chitose (Mrs. Ito Oda), and no study has verified that the same 'self / original speaker' reading is obligatory in other dialects Bugaeva (2008: 42–43). Second, the distributional boundary between the genre-bound narration convention (§5) and a grammaticalized logophoric device is not resolved: quantified evidence of asinuma in everyday versus oral-literary speech is not available in the sources consulted ‹speculative›. Third, the non-Saru dialect behaviour of the quotative system — what role ari/ani plays in the DRD/SDRD/IRD split, and whether the same transitivity constraint holds — has not been described at the same level of detail as the Saru–Chitose baseline in Bugaeva (2008) and Tamura (2000).