Chapter 77The Permissive/Sociative Causative -yar
The indirect/permissive causative -yar 'let/have someone do' and the directive-vs-sociative split it forms with -re/-e/-te.
77.1 Place in the causative system
Hokkaido Ainu derives causative verbs by two principal suffix types that differ in their effect on the argument frame. The productive causative -re/-te/-e, treated in Chapter 74 (The Morphological Causative -re / -e / -te), raises verb arity by one: a new causer takes the subject position and the original subject is demoted to object. The suffix -yar, and its allomorph -ar, introduces a causer as the new subject but suppresses the causee entirely — the party that carries out the caused action is neither expressed as a separate argument nor cross-referenced on the verb, so the net argument count of the derived form equals that of its base Nakagawa (2024: 201–202); Bugaeva & Kobayashi (2022: §4.3.1); Kobayashi (2015: 12). Nakagawa's label for this type is 不定使役 'indefinite causative' Nakagawa (2024: 201); the Handbook chapter on verbal valency places it among the valency-neutral alternations alongside the sociative uko-/uwe- Bugaeva & Kobayashi (2022: §4.3.1). The broader arity calculus that frames all valency operations is set out in the overview chapter; causee marking and double causatives are in Chapter 76 (Causee Marking, Ditransitive and Double Causatives).
| type | suffix | arity | causee | label |
|---|---|---|---|---|
| productive causative | -re / -te / -e | +1 | expressed as direct object; cross-referenced by a person index | 使役 |
| indefinite causative | -yar / -ar | 0 | unexpressed; cannot be cross-referenced | 不定使役 |
A causative suffix that leaves argument count unchanged is typologically unusual Nakagawa (2024: 201–202); Bugaeva & Kobayashi (2022: §4.3.1). Within the arity system the outcome is coherent: -yar encodes a new causal event while the causee is structurally absorbed, in the same way that noun incorporation satisfies an argument slot from inside the verb word without expanding external valency Nakagawa (2024: 201).
77.2 Form and allomorphy
The suffix surfaces as -yar after a vowel or a glide and as -ar after other consonants Bugaeva & Kobayashi (2022: §4.3.1); Nakagawa (2024: 202). The conditioning environment mirrors that of the productive causative, where vowel- and glide-final stems take -re, most consonant-final stems take -te, and r-final stems take -e Nakagawa (2024: 200–201). The canonical -yar form is nu ‘hear (something)’ → nu-yar 'make someone unspecified hear something, tell'. The allomorph -ar would surface on consonant-final bases such as a hypothetical sinot-ar 'have someone play'; the allomorphy is stated in the Handbook but no consonant-final form is attested in the read corpus sources Bugaeva & Kobayashi (2022: §4.3.1) ‹corpus-suggested›.
Nakagawa's slot table places -yar/-ar in the causative suffix position, after the stem and any forming material and before the suffixal person index Nakagawa (2024: 200, 210–211). Any suffixal person index on the derived form marks the new causer-subject, not the suppressed causee.
77.3 The unexpressed causee
The causee — the party who carries out the caused action — is structurally absent from the derived verb: no person index cross-references it and no bare NP expresses it as a core argument. Satō states that the suffix cannot license a separate causee NP; Kobayashi's valency study describes the type as 不定使役 with the causee role suppressed and valency unchanged Satō (2008: 234); Kobayashi (2015: 12). This separates -yar from the productive causative, which expresses the causee as a direct object, cross-references it with a person index, and produces a double-object frame when the base is already transitive Nakagawa (2024: 201–202).
When a speaker wishes to indicate the implied original agent despite its exclusion from the verb's argument frame, the postpositions or peka, or un, or eun supply a peripheral adjunct slot. Nakagawa's Saru examples (exx. 961–963) include nispa or peka a=korar 'marry off to a rich man', where or peka specifies the intended recipient without making that recipient a core verbal argument Nakagawa (2024: 429). The same option appears in oral narrative, where a deity named in an eun phrase receives the message conveyed by nu-yar while the verb itself indexes no causee Biratori Ainu oral literature (1969).
‘I felt I couldn't have her married off far away.’
Biratori Ainu oral literature 1969; Saru, Biratori oral lit., 木村きみ narration
korar (kor-ar, 'marry off') is the indefinite causative: the causer a= (4th-person, logophoric, referring to the narrator) is indexed on the verb, but the causee — the woman married off — is neither expressed nor cross-referenced, recoverable only from discourse. This is the structural signature of -ar.
‘I had it heard / I reported (it) to [them].’
Biratori Ainu oral literature 1969; Saru, 貝澤 とぅるしの narration (biratori/004/002#80)
nuyar (nu-yar) keeps the two-place frame of nu 'hear something': the causer a= (4th-person, logophoric) is indexed on the verb, while the recipient of the message is supplied by the eun adjunct and the causee (the hearer) is neither expressed nor cross-referenced — the structural signature of the indefinite causative.
77.4 Readings: indefinite, permissive, and sociative
Three semantic characterisations appear in or near the literature, and the sources consulted converge only on the structural valency facts while leaving the label question open Nakagawa (2024: 201–202); Bugaeva & Kobayashi (2022: §4.3.1) ‹contested›.
The indefinite-causation reading — the causer causes an event but the party who performs it is unspecified (誰にでも 'to whoever') — is the characterisation that Nakagawa and the Handbook share Nakagawa (2024: 201–202); Bugaeva & Kobayashi (2022: §4.3.1). It fits the most frequent attested form: nu-yar functions as the regular verb for public announcement or reporting, a speech act whose point is precisely that the recipient class is left open.
‘Everything I was asked I told in full — I made it all known.’
Nakagawa 2024: 432; Saru, Nakagawa ex. 959, after Tamura 1984: 12
ku=nu-yar retains the two-place transitivity of nu 'hear something'; the content made known is the sole object and no causee argument is present.
The permissive reading surfaces in contexts of polite or deferential causation, where the causer enables an action that the unnamed causee has reason or inclination to perform. Bugaeva and Kobayashi note that a host may say ipe-yar=an ro 'let's have the guest eat' rather than naming the guest as a direct causee, producing a softer formulation Bugaeva & Kobayashi (2022: §4.3.1); Tamura (2000: 214). Bugaeva's Saru study shows that the productive -re/-te/-e also covers permissive causation alongside coercive causation, and that ingestive verbs like 'eat' and 'drink' take both readings with that suffix Bugaeva (2014: 139, 141–142). The structural property of -yar — suppression of the causee — is what distinguishes it from -re/-te/-e in permissive contexts; the permissive vs coercive axis alone does not predict the choice between the two suffix types.
‘Let's have (them) eat.’
Bugaeva & Kobayashi 2022: §4.3.1; Saru
After Tamura 2000[1988]: 214. The intransitive S suffix =an on ipe-yar confirms that the derived verb remains intransitive; the eating party is implied by context but not indexed on the verb.
The label sociative causative — marking the causer's participation together with or on behalf of others — has been mentioned in connection with -yar alongside the indefinite and permissive labels. The sources consulted do not adjudicate between the three characterisations. Bugaeva's 2014 causative study, the most comprehensive treatment of Ainu causative morphology available, explicitly excludes -yar for reasons of space Bugaeva (2014: 127).
77.5 Contrast with the productive causative
| parameter | -re / -te / -e | -yar / -ar |
|---|---|---|
| arity | +1 (causer added; old subject → object) | 0 (causer added; old subject suppressed) |
| causee expression | direct object; person-indexed | unexpressed; no person indexing |
| specifying the implied agent | — | or peka / or un / eun adjunct |
| causation type | direct and indirect | indirect only |
| productivity | high (190 forms in Bugaeva's Saru sample) | low; few attested forms |
The data for this table come from Nakagawa (2024: 200–202), Bugaeva (2014: 127, 130, 143), and Bugaeva & Kobayashi (2022: §4.3.1).
The pair nu-re / nu-yar makes the contrast explicit. The productive causative nu-re 'make someone hear something' is a three-place verb: the causer is subject, the causee occupies one object slot and receives a person index, and the content heard fills the second object slot. The indefinite causative nu-yar keeps the two-place frame of the base nu ‘hear (something)’: the causer is subject, the content is the sole object, and the recipient is absent from the verb's morphosyntax Nakagawa (2024: 201–202).
77.6 Corpus profile
Frequency data from the Hokkaido corpus show the suffix active in only a small set of forms. The token ipeyar returns approximately 2 hits (rank 14,619), making it marginal; nuyar is more frequent but the absolute total remains small ‹corpus-confirmed›. The distribution reflects lexicalized or semi-lexicalized combinations rather than productive attachment across the verb lexicon Bugaeva & Kobayashi (2022: §4.3.1).
nu-yar's higher frequency follows from its speech-act function. The indefinite-causee structure is suited to public announcement: a causer who makes something known without restricting who will receive the information has systematic recourse to this form wherever the open recipient set is the point. The permissive-polite use — ipe-yar=an ro 'let's have the guest eat' — is attested but at low frequency ‹corpus-confirmed›.
Ijäs's learner grammar does not treat -yar; Lesson 9 on special verb forms discusses the causative, reciprocal, and antipassive but covers only -re among the causative types Ijäs (2023).
77.7 Open questions
Several questions are unanswered in the sources consulted. Bugaeva's 387-verb Saru causative study — the most comprehensive available — excludes -yar for reasons of space Bugaeva (2014: 127). No source in the present corpus offers a distributional account of which verb classes the suffix selects, what semantic or pragmatic factors predict the choice between -yar and -re/-te/-e on the same base, or how the suffix patterns dialectally outside Saru. The structural facts — arity 0, causee unexpressed, or peka/or un/eun adjunct option for the implied agent — are consistent across the four sources that address the suffix Nakagawa (2024: 201–202, 429); Bugaeva & Kobayashi (2022: §4.3.1); Kobayashi (2015: 12); Satō (2008: 234). The semantic label — whether indefinite, permissive, or sociative causation best characterises the suffix — remains open ‹contested›.
One diachronic proposal treats -yar as descending from a pronominal-causative construction *i=a-de, with regular intervocalic *d > r; the zero-valency outcome would follow from the prior grammaticalization of the indefinite causee pronoun into the verb complex, parallel to how the meaning of i=…-re semantically overlaps with that of -yar (aynu-corpora Discord (2023–2026); antitwilight, aynu-corpora Discord, 26 May 2026) ‹speculative›. No peer-reviewed source develops this account.
References cited in this chapter
aynu-corpora Discord (2023–2026) ·Biratori Ainu oral literature (1969) ·Bugaeva (2014) ·Bugaeva & Kobayashi (2022) ·Ijäs (2023) ·Kobayashi (2015) ·Nakagawa (2024) ·Satō (2008) ·Tamura (2000)