Chapter 105Quotative Complementation and Reported Speech (sekor)

105.1 sekor in the complementizer system

Hokkaido Ainu reports speech and thought by placing the invariable particle sekor at the right edge of the quoted string, before the matrix verb of saying or thinking. This mechanism occupies a distinct slot in Refsing's four-way clause-embedding typology: conjunctionalization, nominalization, quotationalization, and adnominalization — with quotationalization, realized by sekor (and its variant ari), forming a separate category from the nominalized-complementizer system Refsing (1986: §14.14). The nominalized complementizers — kuni, kunak, yak, kusu — that Nakagawa treats in his indirect-speech section operate through the formal-noun nominalizing machinery described in Chapter 101 (The Nominalizers -p/-pe and -i/-hi (Participant, Event, Place, Fact)) and Chapter 106 (Nominalized Complements and Control/Raising); sekor does not nominalize — it quotes directly Nakagawa (2024: 475–480).

The matrix verbs that take sekor overlap substantially with those that select the indirect-speech markers: ye ‘say’, itak ‘speak, say’, hawean ‘say (singular subject)’, and ramu ‘think’. The choice between the quotative and the nominalizing complementizers encodes the direct-vs-indirect distinction and controls the person-marking pattern in the embedded clause Nakagawa (2024: 475); Bugaeva (2008: 31–35). How the formal-noun class that enables nominalized complementation is constituted is set out in Chapter 35 (The Formal / Defective Noun Set (ruwe, hawe, siri, humi, hi, pe, kur)).

105.2 Form and position

sekor is morphologically invariant: it carries no person affixes, no TAM morphology, and no allomorphic alternation conditioned by the shape of the preceding clause. Its canonical position is immediately after the last element of the quoted string and before the matrix verb: [quoted clause] sekor [matrix verb]. Refsing identifies a variant ari with the same quotative function in the Shizunai dialect; both are treated as "quotationalizers" forming a class distinct from the nominalized-complementizer inventory Refsing (1986: §14.14).

The quoted clause preserves all features of a main clause: person affixes, polarity, TAM markers, and clause-internal particles are intact inside the quotation. Sentence-final particles (na, wa, etc.) may appear within the quoted string but not between sekor and the matrix verb — a distributional restriction shared with the indirect-speech markers, whose embedded clauses also exclude final particles Nakagawa (2024: 475–476).

105.3 Direct and indirect quotation

The clearest evidence for the sekor vs indirect-complementizer distinction is person marking. Under sekor, the original speaker's person forms are preserved unchanged inside the quote. Under kuni, kunak, and yak, the embedded clause's person is adjusted to the reporting narrator's viewpoint: an original first person shifts to the fourth, the indefinite/logophoric person Nakagawa (2024: 475). The indirect-speech markers also differ among themselves in the temporal or semantic value they impose on the embedded proposition Nakagawa (2024: 475–480); Satō (2008: 176–177).

The person contrast between sekor and kuni is illustrated by the following pair. The proposition 'I will return' appears first as a direct quotation (constructed), then under the indirect-speech complementizer kuni (attested in Nakagawa's grammar):

The temporal contrast between kunak (future proposition, reported as ongoing) and yak (past or general proposition) runs in parallel within the indirect system. Satō's kunak example combines the indirect-speech marker with the ongoing-aspect phrase kor an:

The directive use of kuni — 'tell X to do Y' — appears frequently in folktale dialogue and marks the complement as a directed action rather than a proposition:

The purposive and deontic extensions of kuni beyond the reported-speech domain are treated in Chapter 107 (The kuni Complementizer and Purpose Complements).

105.4 hawe and itak as framing elements

The formal noun hawe — possessed form of haw ‘voice’, 'the voice/report of' — extends the quotative architecture in two directions. As a bare complementizer it nominalizes a speech event and supplies it as the perceptual object of verbs like nu ‘hear, learn of’: the complement clause becomes "the voice/talk of" the event Satō (2008: 175–178). Combined with the copula, hawe ne forms the reportative evidential frame; when sekor precedes it, the compound sekor hawe ne attributes a quoted proposition to an unspecified external source ‹corpus-suggested›:

The clause-final reportative evidential hawe ne standing alone — where it marks information acquired through speech without framing an embedded quotation — is treated in Chapter 123 (hawe ne — the Reportative / Hearsay Evidential), and the four-way evidential paradigm to which it belongs in Chapter 120 (The Nominalization-plus-Copula Evidential Schema).

itak ‘speak, say, talk’ is the most frequent speech verb in narrative dialogue — narrators routinely use it to introduce character utterances — while ye ‘say’ appears predominantly in indirect-speech and complementizer frames. Both appear after sekor and both select kuni/kunak for indirect complements Nakagawa (2024: 476); Satō (2008: 176). A stable semantic or register distinction between the two verbs in quotative syntax is not established in the sources consulted ‹corpus-suggested›.

105.5 The logophoric fourth person in speech reports

Person adjustment in the indirect-speech complementizer frames reflects what Bugaeva identifies as logophoricity in Southern Hokkaido Ainu. The fourth-person markers a=/ci= on transitives and =an on intransitives replace an original first person in the embedded clause, reanchoring it to the reporting narrator's viewpoint. Nakagawa formulates this as a categorical rule: under the indirect-speech markers, "final particles cannot be embedded; the speaker's-viewpoint person shifts 1→3/4" Nakagawa (2024: 475).

Bugaeva's study of reported discourse and logophoricity in Southern Hokkaido Ainu provides the analytical framework: the fourth person here is not a simple third-person referential shift but a logophoric device, indexing the embedded protagonist — the original speaker or thinker — whose perspective the narrator temporarily adopts Bugaeva (2008). The surface form is identical to the generic/indefinite fourth person (also a=/=an); the logophoric reading is determined by the presence of a complementizer frame and by contextual co-reference with the matrix subject ‹contested›. The full fourth-person paradigm and its logophoric distribution across narrative genres are treated in Chapter 68 (Honorific and Logophoric Uses of the Fourth Person) and Chapter 67 (The Indefinite/Fourth Person: Forms, Reference, and Agent-Defocusing).

Nikitina and Bugaeva develop a stronger claim: the construction is better characterized as logophoric speech rather than indirect speech in the conventional theoretical sense, because the person system encodes the narrator's perspective adoption, not a syntactic binding of the embedded subject to a matrix controller Nikitina & Bugaeva (2021). The discourse contexts — first-person narration in uwepeker and yukar — where this perspective adoption is most systematic are examined in Chapter 150 (First-Person Narration, the Logophoric Fourth Person, and Reported Discourse) ‹contested›.

Under sekor direct quotation, no person shift applies: the original speaker's person forms are preserved inside the quote, and the logophoric reanchoring does not arise in the embedded clause. This is the morphological correlate of the direct/indirect boundary: sekor preserves the perspective of the original utterance; the indirect-speech markers reanchor it to the narrator's viewpoint Bugaeva (2008: 31–35); Nikitina & Bugaeva (2021).

105.6 Analyses and open questions

The descriptions converge on the categorical point. Refsing separates quotationalization from nominalization as coordinate clause-embedding types, placing sekor/ari outside the nominalizer class Refsing (1986: §14.14). Nakagawa's §18.2, which treats the indirect-speech markers extensively as nominalization-based complementizers that take a nominal-object slot, is consistent with keeping sekor as a non-nominalizing quotative; his account of the person shift and the final-particle ban applies to kuni/kunak/yak and not to sekor Nakagawa (2024: 475–480). Shibatani, whose treatment foregrounds the formal-noun complementizer system, does not single out a quotative particle in the dialect data he covers Shibatani (1990: 83–84).

Three questions remain open in the sources consulted. First, the geographical and register distribution of sekor vs ari across Hokkaido dialects is documented only by Refsing's Shizunai attestation of both Refsing (1986: §14.14); whether ari carries any semantic or pragmatic distinction from sekor or is a register/lexical variant is not established ‹speculative›. Second, the grammaticalization status of the compound sekor hawe ne frame is not settled: whether the sequence is transparently compositional (quotative + reportative evidential) or a lexicalized reportative formula cannot be decided from the sources read here ‹speculative›. Third, the logophoricity analysis of Nikitina and Bugaeva awaits corpus-level testing at construction level; whether the person-shift pattern under kunak/yak distributes differently from the generic fourth-person use is not settled by evidence in the sources consulted ‹contested›.

Ijäs's pedagogical grammar covers relative clauses and the formal-noun nominalizer system in Lesson 11 — the material banked for this grammar's nominalization domain — but her treatment of direct and indirect speech falls in a separate lesson not yet incorporated here Ijäs (2023). No Ijäs claims on sekor or quotative framing are available for grading here.