Chapter 133The hine/akusu Contrast and Switch-Reference-Like Linkage
The sequential connectives hine and akusu and the long-running debate over whether they encode same-/different-subject (switch-reference) or discourse (un)expectedness.
133.1 The two particles and the question they raise
Two conjunctive particles anchor this chapter. hine ‘and then (neutral chainer)’ closes a finite clause and places it before the next in temporal order, without committing to any causal, manner, or integrative relation between the two events. akusu ‘and then (discovery)’ closes a realized first event and opens the clause in which a consequence is encountered or witnessed. Both are bound, post-predicate particles — not verbal suffixes, not converbs — and both join fully finite clauses that carry their own person indexing, polarity, aspect, and any evidential tail, in the pattern established for the whole inventory in Chapter 131 (Clause Linkage in Ainu: Typology and Inventory). The typological question they raise — whether their distribution tracks same-subject versus different-subject reference in the manner of switch-reference morphology (Foley & Van Valin 1984) — runs through the descriptive literature on Ainu clause-linking. The evidence supports a discourse-semantic account rather than a grammaticalized subject-tracking one: neither particle carries a same-subject or different-subject feature, and the observed correlations with subject continuity follow from the particles' discourse-semantic roles rather than from an independent reference-tracking function.
133.2 hine: neutral temporal enumeration
hine places two events in temporal sequence and takes no stance on how they are related. Nakagawa's characterization is precise: hine "lists events in time order" without any commitment to causal, manner, or purpose connection Nakagawa (2024: 338). The bond is therefore looser than wa ‘and (integrative chainer)’, which fuses two acts into one coherent event — presenting the first as the method, result-state, or cause of the second — and looser than akusu, which signals a discovered or witnessed consequence Nakagawa (2024: 338–339); Satō (2008: 44); Refsing (1986: 242).
In Saru and Chitose prose, hine is the dominant chainer for the stock opening that establishes the narrator's household before the story begins. The formula … an hine oka=an — one or more existence clauses strung together and closed by the plural existential 'we lived on' — recurs with minimal variation across the genre. The pattern is so regular that Refsing identifies hine as more frequent in folk tales than in colloquial speech Refsing (1986: 242):
‘My father was there, my mother was there, and we lived on.’
Nakagawa 2024: 338; Saru
Nakagawa ex. 601; a stock prose-tale opening. The three items — father, mother, living — are listed in order with hine as a pure temporal enumerator, no causal claim made.
Corpus texts show the same frame pivoting into the episode proper via the formula oka=an pe ne hike 'and we lived on, and …', where the formal-noun connective hike ‘and (loose link)’ opens the narrative event:
‘My elder sister in the sky raised me, and we lived on, and …’
ILCAA Ainu materials 1976; Saru, 川上まつ子「uwepeker 民話15」, AA研アイヌ語資料, aa-irc/022#0
The hine chain closes on oka=an pe ne hike, the standard narrative-pivot formula. The formal-noun pe and the loose-link hike then open the episode proper.
133.2.1 Etymology, phonology, and dialect distribution
The particle derives from the nominalizer hi ‘the fact that; the thing that’ plus the copula ne ‘be, is’: hi + ne > hine (Tamura 1972b: 158, via Refsing (1986: 243); Kindaichi 1960: 193, via Refsing (1986: 243)). The source material is reached here through intermediaries; the etymology is reported, not independently verified. It aligns with a wider pattern in which nominalization-plus-copula material grammaticalizes into clause-linkers, visible also in the evidential system (see Chapter 121 (ruwe ne — the Inferential / Visual-Trace Evidential)).
Phonologically, word-initial h is regularly dropped in normal-speed speech, and the resulting vowel i fuses with a preceding final consonant: ek hine 'came-and' → [ekine]; a=ne hine → [aneine]. Ishikari, where initial h is generally absent, writes the form ine. In rapid speech the vowel itself may elide to bare [ne] Nakagawa (2024: 338); Refsing (1986: 243).
Dialect distribution is uneven. Tokachi uses hinne (e.g. Honbetsu) and hine in alternation; Shiranuka and Kushiro write ine; Honbetsu lacks both forms entirely, the chaining function falling to other resources. In Shiranuka, frequency is lower than wa and tek combined Nakagawa (2024: 338). The whole-corpus count of approximately 14,368 tokens — the highest-frequency temporal-sequence particle with few competing homonyms — is a usable signal of the particle's prominence, with the usual caveat that Hokkaido and Sakhalin varieties are not separated in the aggregated count ‹corpus-suggested›.
133.3 akusu and awa: sequential discovery
akusu and awa belong to Nakagawa's group-② 'unfolding event' (事態の展開) connectives, alongside hike and ayne ‘and finally (after a protracted action)’ Nakagawa (2024: 345–347). Their core function is to present a completed first event followed by a second event that comes to the participant's attention — "the anterior event actually happened, and then coincidentally (偶発的に) the posterior event occurred" Satō (2009: 52). Satō classes the construction as a factual condition (事実的条件): the protasis is realized rather than hypothetical, and the apodosis reports the encountered consequence Satō (2008: 47); Satō (2009: 52).
The discovery flavor of akusu — English 'whereupon', 'and then (I found/saw that)' — is not a causal claim. Nakagawa makes this explicit: the causal connection between E1 and E2 is irrelevant to the particle; what it marks is that E2 comes to the attention of a participant after E1 is complete, whatever the causal structure may be Nakagawa (2024: 345):
‘After killing that small deer, …’
ILCAA Ainu materials 1976; Saru, 川上まつ子「isoytak 独話1」, AA研アイヌ語資料, aa-irc/003#73
The fourth-person agent a= completes the killing; akusu opens the clause in which the narrator discovers or experiences what follows. The continuation is not preserved in the banked fragment.
Akusu attaches after a clause-final evidential tail — confirming the ordering [VP]–evidential–connective established for the system as a whole. When the first clause closes on ruwe ne, the anterior event is presented as an established fact before the discovery linker opens the next clause:
‘I had sent a letter, and in return a letter came.’
Satō 2008: 47; Satō 2009: 52; Chitose
Satō ex. 16 / Satō 2009 ex. 32. The discovery linker attaches outside the ruwe ne evidential tail. The cross-linguistic asymmetry with Japanese — factual と/たら cannot follow preceding のだ, but Ainu akusu can follow ruwe ne — is discussed in §5.
133.3.1 The akusu/awa register and dialect split
Akusu and awa are semantically equivalent; no descriptive source establishes any difference in meaning. Their distribution is conditioned by register and dialect: in Saru, prose texts favour akusu while verse texts heavily favour awa; Horobetsu and Ishikari use awa regardless of register; Shizunai and Tokachi use akusu Nakagawa (2024: 345). Satō's characterization of awa as the oral-literature (口承文芸) variant aligns with the verse skew in Saru Satō (2008: 47–48) ‹consensus›:
| particle | Saru prose | Saru verse | Horobetsu / Ishikari | Shizunai / Tokachi | etymology |
|---|---|---|---|---|---|
| akusu / akus | primary | rare | rare | primary | a (PFV) + kusu (consequential) |
| awa | rare | primary | primary | rare | a (PFV) + ? |
The etymology of akusu as a (perfective auxiliary) + kusu (consequential) is supported by Refsing and is consistent with the particle's semantics — a completed event (a: 'situation done and closed') from which a consequence follows (kusu) Refsing (1986: 248). The second morpheme of awa has not been identified with confidence Refsing (1986: 248). Shizunai uses the phonologically reduced form akus (word-final vowel devoicing and deletion) Refsing (1986: 248).
Whole-corpus word-form counts register approximately 3,861 tokens for akusu and 2,071 for awa. Both are among the cleaner particle counts with few homonyms, and the ratio is consistent with a mixed-register corpus containing proportionally more prose than verse ‹corpus-suggested›. Per-function counts that would separate Hokkaido from Sakhalin are not available from the current search tools.
133.4 The switch-reference question
Switch-reference — dedicated morphological marking on clause-linkers that co-indexes subject continuity between clauses (Foley & Van Valin 1984) — is absent from Hokkaido Ainu. The language has no converbs, no non-finite verb forms, and no SS/DS feature on any conjunctive particle Bugaeva (2022); Nakagawa (2024: 514). What can be asked is whether the choice among particles statistically or semantically correlates with subject continuity, and whether any such correlation is grammaticalized or follows from the particles' discourse-semantic content.
133.4.1 Statistical tendencies and their source
The one quantified figure in the literature is Refsing's Shizunai count: approximately ninety per cent of the predicate pairs joined by wa share a subject Refsing (1986: 240). No comparable count exists for hine or akusu in any Hokkaido variety; the quantitative base is a single informant's Shizunai material. The ninety-per-cent figure is consistent with a preference but demonstrably not with a constraint: Satō shows that wa tolerates different-subject junctions without any change in form, as in the child-throws-stone / barrel-falls example where the thrower and the barrel are different participants — a junction Japanese 〜て cannot make Satō (2008: 45–46). No primary grammar states a same-subject requirement for hine, and its neutral-listing semantics entails no such restriction Nakagawa (2024: 338); Refsing (1986: 242).
The simultaneity particle kor ‘while (simultaneous)’ provides a direct test: it takes different-subject complements without restriction, as in Saru ku=horipi kor en=nukar 'she looked at me while I danced' Ijäs (2023), corroborated by an oral-literature example in Satō Satō (2008: 49–50). Shizunai kane "generally" takes same-subject complements but exceptions are attested Refsing (1986: 250). Across the particle inventory, any same-subject preference is a graded tendency, not a binary SS/DS grammatical feature.
Ijäs offers the pedagogical heuristic that negated clauses take no when the two predicates share a subject and wa when they do not Ijäs (2023). Nakagawa and Satō derive the no/wa choice from the aspect and polarity of the first predicate — stative and negated action predicates take no — rather than from subject identity Nakagawa (2024: 340–341); Satō (2008: 45–47). The Ijäs framing is noted because it has entered pedagogical circulation ‹contested›; the primary descriptive sources do not support it.
133.4.2 Expectedness and discourse stance
Nakagawa's and Satō's analyses locate the hine / akusu contrast in event structure and narrative stance, not in subject continuity. The neutral temporal enumerator hine naturally chains a single protagonist's sequential acts — producing a predominance of same-subject pairs without encoding subject identity. The discovery connective akusu marks that the narrator encounters or witnesses the second event after the first is complete, a framing that often foregrounds a new or unexpected participant and thereby produces an apparent different-subject tendency. The observed SS/DS distributions are statistical consequences of discourse function, not evidence for a grammaticalized opposition Nakagawa (2024: 338, 345); Satō (2008: 44, 47) ‹consensus› (for the basic semantic descriptions).
133.4.3 The narrative boundary device
The nearest structural analogue to a switch-reference signal is discourse-organizational rather than morphological. The Handbook's descriptive grammar notes a tendency for the sequence of the perfective auxiliary a ‘perfective (situation complete)’ followed by a connective such as korka ‘but, although’, wa, or the formal-noun connective p ‘thing (formal noun as connective)’ to fall at transitions of scene, topic, or acting participant in narrative Bugaeva (2022) ‹speculative›. The perfective marks the anterior situation as completed and cut off from the present; an adverb of time typically follows to open the new episode. The printed location of this observation has not been confirmed against the Handbook print edition, so the claim is treated as speculative.
The device is a clause-chaining boundary convention. No morpheme on the connective particle encodes subject identity; the signal comes from the co-occurrence of the perfective auxiliary with a particular linker at a discourse transition. Subject identity across a chain is tracked by the obligatory person affixes on each finite predicate, the mechanism drawn together in Chapter 139 (Reference Tracking and Argument Continuity).
133.4.4 Graded positions
The competing positions, arranged by evidence grade:
| position | source | grade |
|---|---|---|
| No morphological switch-reference; conjunctive particles carry no SS/DS feature; person affixes track argument continuity | Bugaeva (2022); Nakagawa (2024: 514) | consensus |
| wa tends same-subject (≈90% in Shizunai) but imposes no grammatical constraint | Refsing (1986: 240); Satō (2008: 45–46) | corpus-suggested (one dialect, one informant; no Saru count for hine) |
| hine / akusu encode discourse stance (neutral ordering vs discovery), not subject continuity | Nakagawa (2024: 338, 345); Satō (2008: 44, 47) | consensus |
| no / wa conditioned by SS vs DS | Ijäs (2023) | contested — primary grammars condition on polarity/aspect, not subject identity |
| Perfective a + korka / p tends to fall at scene and actor transitions | Bugaeva (2022) (Handbook Part II, page unverified) | speculative |
| The SS tendency of hine and the apparent DS lean of akusu both follow from discourse semantics; no separate reference-tracking feature is warranted; corpus annotation of subject continuity at these junctions would provide the empirical test | synthesis from the above; no single source states this directly | original-needs-review |
The ‹speculative› synthesis draws on what each source establishes independently. Hine neutrally chains events along a single narrative strand — which by default is the protagonist's sequential acts — hence the same-subject predominance without a grammatical constraint. Akusu frames the second event as encountered or discovered after the first is complete, a perspective that frequently involves a new or surprising participant and thereby shows an apparent different-subject lean. Neither particle requires a particular subject configuration; the SS/DS distributions reflect the default narrative configuration of each connective's domain. A corpus annotation of subject continuity across hine and akusu junctions in Saru text would supply the empirical test ‹speculative›.
133.5 The evidential–connective interface
Conjunctive particles attach outside the clause-final evidential frame. The ordering [VP]–evidential–connective, established for the system as a whole in Chapter 131 (Clause Linkage in Ainu: Typology and Inventory), means that akusu can follow a ruwe ne tail, yielding … ruwe ne akusu …. The evidential sits above clause-internal aspect but below the clause-linker, so kusu ne ruwe ne is grammatical and *ruwe ne kusu ne is not Nakagawa (2024: 393).
Satō identifies a typologically notable property of this interface. Japanese factual conditional と / たら cannot follow a preceding のだ, yet Ainu places no such constraint on the ruwe ne akusu sequence: the first event is presented as knowledge-established fact and the discovery linker then opens the encountered consequence Satō (2009: 52–53). The construction presents the two layers as separately stackable because they occupy distinct positions in the clausal hierarchy.
A member of the aynu-corpora Discord community corroborates both the frequency and the discourse function of this sequence. nukopoli (aynu-corpora Discord, 2024) observes that ruwe ne is associated with scene transition and summary framing in prose narrative, and that ruwe ne akusu functions as a compound of that discourse anchoring with the discovery connective aynu-corpora Discord (2023–2026) ‹speculative› — corroborated by Satō (2009: 52). The observation is consistent with the evidential+linker ordering described by Nakagawa Nakagawa (2024: 393) and with the ruwe ne system set out in Chapter 121 (ruwe ne — the Inferential / Visual-Trace Evidential).
133.6 Dialect edges
The particles described above rest primarily on Saru and Chitose attestation. Several dialect-specific details bear on the comparison. Shizunai uses akus (the word-final vowel devoiced and deleted) for the same discovery function Refsing (1986: 248). Horobetsu and Ishikari favour awa regardless of register, while prose-oriented Shizunai and Tokachi use akusu as the default Nakagawa (2024: 345). The category is stable across the dialect continuum; phonological shape and genre weighting shift, not the underlying semantic function.
Tokachi uses hinne alongside hine for neutral chaining; Honbetsu Tokachi shows neither form. These micro-variations caution against projecting Saru frequency data onto the dialect cluster as a whole Nakagawa (2024: 338). The full topology of the conjunctive-particle system across Hokkaido dialects is surveyed in Chapter 131 (Clause Linkage in Ainu: Typology and Inventory); the conditional and concessive members of the system are treated in Chapter 135 (Conditional, Temporal, and Concessive Clauses).
133.7 Open questions
The foundational Saru description of the conjunctive particles, Tamura's 1972 study of the 〜して constructions (Tamura 1972b), is the primary source for nearly every claim about hine in the grammars consulted here, but it was not read directly — all attributed page numbers reach it through Nakagawa (2024) and Refsing (1986) ‹speculative› (the page attributions are second-hand). The quantitative base is thin: Refsing's ninety-per-cent same-subject count for wa in Shizunai is the only published figure; no per-function counts for hine or akusu in any Hokkaido variety exist in the accessible literature. The whole-corpus totals — approximately 14,368 for hine, 3,861 for akusu, 2,071 for awa — are homonym-mixed, Hokkaido-and-Sakhalin combined, and carry no subject-continuity annotation. Whether akusu junctions show measurably lower same-subject rates than hine junctions — the empirical test of the ‹speculative› synthesis — remains open and requires bespoke corpus annotation over a tagged Saru text. The reference-tracking apparatus as a whole is drawn together in Chapter 139 (Reference Tracking and Argument Continuity); narrative cohesion and tail-head linkage are treated in Chapter 138 (Tail-Head Linkage and Narrative Clause Cohesion).
References cited in this chapter
aynu-corpora Discord (2023–2026) ·Bugaeva (2022) ·Foley & Van Valin (1984) ·Ijäs (2023) ·ILCAA Ainu materials (1976) ·Nakagawa (2024) ·Refsing (1986) ·Satō (2008) ·Satō (2009)