Chapter 74The Morphological Causative -re / -e / -te

74.1 Place in the causative inventory

Hokkaido Ainu marks causation with a small set of verb-final suffixes. The productive majority is one underlying form realized as three phonologically conditioned allomorphs: -re after a vowel or the glides y, w; -te after any consonant other than r; and -e after r Nakagawa (2024: 200–201); Bugaeva (2014: 134). In Nakagawa's arity calculus it carries the value +1: the original subject of the base verb is reassigned to the object position of the derived verb, and a new causer participant fills the subject position Nakagawa (2024: 200–201, 431).

Three further suffix types occupy the same post-stem slot: the forming suffixes -V and -ke that build transitivity pairs, the direct-causative / transitivizer -ka, and the valency-neutral indefinite causative -yar / -ar. The indefinite causative, which leaves the causee unexpressed and unindexed, is treated in Chapter 77 (The Permissive/Sociative Causative -yar); the forming suffixes and -ka are covered in Chapter 75 (The Transitivizer -ka and Lexical Valency Pairs). Causee marking, the double-object frame, and double causatives are developed in Chapter 76 (Causee Marking, Ditransitive and Double Causatives). The present chapter concerns -re / -te / -e alone: their allomorphy, their arity effect, the productive versus non-productive split, and the direct/indirect and coercive/permissive distinctions.

74.2 Allomorphy

The three surface forms are phonologically predictable from the stem-final segment Nakagawa (2024: 201); Bugaeva (2014: 134); (Tamura 2000). Tamura's Saru dictionary and Bugaeva's 387-verb Saru/HSW corpus sample converge on the same conditioning environments. Satō's Chitose textbook gives the same three allomorphs with the qualification "大体 by stem-final sound, but with exceptions", acknowledging a residue of forms that resist the phonological prediction Satō (2008: 234–238).

One phonological property separates the productive suffix from the non-productive cluster: it does not shift the pitch-accent kernel of the stem — kú-re and sinót-te, not *ku-ré or *sinot-té Bugaeva (2014: 130–131).

The analysis of siknu-re 'keep alive' — intransitive siknu ‘live, be alive’ plus -re — is corroborated in community linguistic discussion aynu-corpora Discord (2023–2026) (nukopoli, aynu-corpora Discord, 2026): the vowel-final stem takes the predicted allomorph without irregularity ‹corpus-confirmed›. A related community observation notes that lexicalized forms such as iyairaykere (< i=yairayke-re, lit. 'makes people grateful') and irankarapte (< i=rankarap-te, lit. 'makes people greet') may carry an "obligatory" or "deserving-of" reading, analogous to participial active/passive contrasts in other languages aynu-corpora Discord (2023–2026) (nukopoli, aynu-corpora Discord, 2024) ‹speculative›. The allomorphic conditioning in these forms follows the regular pattern; the proposed extended reading of the construction is not confirmed in the primary literature.

74.3 Arity and the argument-frame shift

On an intransitive base the productive causative yields a transitive verb: the base's sole argument (S) is reassigned to the object position (O), and the new causer argument fills the subject position (A). The person-marking diagnostics of Chapter 62 (Architecture of the Personal-Affix System: The Four Persons and the S/A/O Paradigms) confirm the output class: a derived transitive takes the prefixal A indexes, while the base intransitive takes suffixal S indexes. Nakagawa positions the causative suffix after the forming and plural material and before the outer person index, placing it in the penultimate layer of the verb-word template Nakagawa (2024: 210–212).

The base-and-derived pair for sinot ‘play’ makes the argument-frame shift visible. In the base clause matkaci 'girl' is the intransitive subject; in the causative clause she is the object (causee), while the causer hapo 'mother' takes the subject position:

When the causee is cross-referenced on the verb by a person index, the argument reassignment is morphologically explicit. Where the first person is the intransitive subject (ku=sinot 'I play'), causativization realigns the same participant as the object en=:

On a transitive base, the productive causative yields a double-object verb (vd): the original A becomes the O causee, the original O is retained as an inner object, and the new causer is the derived A. The attested form a=e=tuy-e-re 'I made you cut (the grapevines)' — tuy-e 'cut' [vt] with -re added — carries two simultaneous person indexes and is cited by Bugaeva and Kobayashi as evidence that causativization of transitives is both licensed and productive Bugaeva & Kobayashi (2022: §4.1.2). Double causatives — the productive suffix added to an already causativized base — are also attested, yielding a four-place verb; the attested kor-e-re 'cause to give' is one such form Bugaeva (2014: 142–143). Both the double-object frame and double causatives are treated in Chapter 76 (Causee Marking, Ditransitive and Double Causatives).

74.4 Productive vs. non-productive causatives

Bugaeva's 387-verb Saru/HSW corpus sample provides the most comprehensive quantitative treatment of the suffix inventory Bugaeva (2014: 130, 143–144). The four suffix types divide by count, base-class, and causation type:

At 190 tokens, the productive suffix outnumbers the three non-productive types combined and is the only one that covers both direct and indirect causation from both vi and vt bases Bugaeva (2014: 143). It also attaches freely to derived stems — those prefixed by applicatives or reciprocal u- — while the non-productive suffixes do not Kobayashi (2015: 33). This combinatorial range marks it as the productive default of Hokkaido Ainu causation ‹consensus›.

Nakagawa separates the forming suffixes -V and -ke as 形成接尾辞 ('forming suffixes') building lexical transitivity pairs, distinct in status from the grammatical causative Nakagawa (2024: 198–203). Bugaeva's analysis subsumes even these under 'direct causative', treating the entire cluster as a valency-increasing operator set with differing scopes Bugaeva (2014: 129–130, 143); Bugaeva & Kobayashi (2022: §4.1.2). The classification of the forming suffixes as lexical versus grammatical causatives is ‹contested›; the relevant forms are developed in Chapter 75 (The Transitivizer -ka and Lexical Valency Pairs).

74.5 Direct and indirect causation; coercive and permissive use

When the base intransitive has an agentive subject, the productive causative yields an indirect causative: the causee acts on its own in response to the causer's enabling, permission, or direction. When the base is stative or non-agentive, the direct reading predominates, with the causee as a highly affected non-volitional patient Bugaeva (2014: 130, 139). The non-productive suffixes are limited to the direct reading; the productive suffix spans both.

An earlier analysis by Fukuda (Tamura) restricted the productive suffix to coercive causation; Bugaeva corrects this, citing Satō 2008, and documents permissive readings from corpus data Bugaeva (2014: 139, 141–142). A verb of ingestion such as e ‘eat’ → e-re carries both coercive ('feed a child') and permissive ('let an adult eat') readings contextually, following the crosslinguistic pattern for ingestive verbs observed by Shibatani and Pardeshi (cited in Bugaeva (2014: 141–142)). Satō 2008 lesson 30 presents the same coercive/permissive ambiguity and derives it from the same form Satō (2008: 237). The following attested example shows the causative of ipe ‘eat a meal’ — ipe-re 'feed' — in a Saru narrative with an indefinite-agent fourth-person prefix:

Ijäs's Lesson 9 formulation that -re "simply turns intransitive verbs into transitive verbs" is a workable approximation for the basic indirect causative; it omits the allomorphic detail, the permissive reading, the causativization of transitives, and the direct-causative uses Ijäs (2023) ‹consensus›.

74.6 Motion verbs and restrictions

A subset of singular motion verbs takes -ke as their primary causative suffix, and Nakagawa documents a semantic split within this class: when a non-volitional object (cargo, luggage) is caused to move, the suffix -ke is standard; when a volitional person is caused to move, the productive -te is used instead Nakagawa (2024: 202–203). The minimal contrast for ahun ‘enter (SG)’ at Nakagawa's examples 196–197:

The -ke / productive-suffix alternation applies to ran ‘descend (SG)’ (ran-ke for objects, ran-te for persons) and yan ‘come ashore (SG)’ (yan-ke for objects) Nakagawa (2024: 202–203). The verb rikin ‘go up (SG)’ takes -ka (rikin-ka) in most dialects; its reduplicated plural allomorph rikip does not feed a causative *rikip-ka, the causative being formed on the singular stem only aynu-corpora Discord (2023–2026) (nukopoli, aynu-corpora Discord, 2026) ‹corpus-confirmed›, consistent with the motion-verb causative paradigm documented by Nakagawa Nakagawa (2024: 202–203).

One restriction spans all causative suffix types: zero-place meteorological verbs resist any causativization, because there is no subject argument to reassign to the object position. sir-pirka 'the weather is good' has no causative *sirpirka-re Bugaeva (2014: 139); Bugaeva & Kobayashi (2022: §4.1.2). Among the non-productive suffixes, -ke additionally requires a one-argument base; the productive suffix and -ka accept both one- and two-argument bases Bugaeva (2014: 139).

74.7 Dialect variation, the -ka → -re/-te shift, and the Horobetsu switch-reference function

The productive suffix is encroaching on -ka territory across Hokkaido dialects, with the shift most advanced in Saru and the southwest. Many verbs that take -ka in northeastern Hokkaido appear with -re/-te/-e in Saru; intra-dialectal doublets such as nam-te ~ nam-ka in Saru document the change in progress Bugaeva (2014: 135–136, 143–144); Kobayashi (2015: 33). Sakhalin varieties (‹SA›) preserve -ka more broadly Bugaeva (2014: 135–136).

The -ka suffix straddles the transitivizer and direct-causative functions in a way the productive suffix does not. Satō 2008 flags some -ke and -e forms as bleached transitivizations — san → san-ke 'put out a fine', where no causer argument is introduced — distinct from the grammatical causative Satō (2008: 237–238). The productive suffix is unambiguously causative in function ‹consensus›.

In the Horobetsu dialect, Takahashi documents an extension of the productive causative suffix into a switch-reference function: in coordinated clauses it marks that the subject of the following predicate differs from that of the preceding one (different subject), while in subordinated clauses it marks same-subject continuity Takahashi (2017: 102–106). The attested Horobetsu forms — é-re 'make eat', kasuy-re 'make help', cis-te 'make cry', wen-te 'make bad' — show the same allomorphic distribution as Saru and Chitose Takahashi (2017: 99–104). Whether the switch-reference function is an independent grammaticalization or a secondary extension of the causative meaning is not settled in the accessible sources; no equivalent use is documented for Saru or Chitose ‹contested›.