Chapter 59Light Verbs, Pro-Verbs, and the Auxiliary-Verb Construction

59.1 The support-verb inventory

Hokkaido Ainu has a small set of high-frequency verbs whose lexical meanings are bleached in particular constructions while their ordinary verbal morphology remains visible. The central member is ki ‘do’, a transitive verb that supplies a verbal host for a preceding event expression; the same support pattern appears in clausal negation, clause linkage under negation, and some pro-verbal uses where the lexical content is recovered from context. The copula ne ‘be, become’ and the existential an ‘exist’ also serve as supports, especially in predicate-nominal, evidential, existential, and aspectual frames Nakagawa (2024: 124–129, 132, 374); Bugaeva (2012: 492–496). The shared condition is structural: the support verb supplies the finite predicate position, so person marking, negation, and clause-linking material attach to a verb word of the kind described in Chapter 56 (The Verb Word and Its Position-Class Template).

The inventory rests on the same valence diagnostics as ordinary verbs. Intransitives take S markers, transitives take A/O architecture, and the side-switching first-plural and fourth-person forms provide the clearest test Nakagawa (2024: 125–128, 165–169); Satō (2008: 94–98). Nakagawa's arity calculus counts person indexes, derivational prefixes, noun incorporation, and base verbs in a single computation; Satō separates argument slots, macrorole transitivity, and morphological valence, using ne as a central case where two argument slots coexist with only one indexable macrorole Nakagawa (2024: 136–139); Satō (2023: 42–52) ‹contested›. The support verbs preserve ordinary head-marking morphology even where their lexical semantics is weak.

59.2 Ki as the formal verb of somo negation

The clearest grammaticalized use of ki is the periphrastic negative frame V (ka) somo ki. The notional predicate precedes the negative adverb somo, optionally followed by the focus/additive particle ka; ki supplies the finite verbal head that somo negates. In Bugaeva's formulation, the preceding notional verb functions as an object-like complement of ki; Nakagawa describes the same configuration as a use of the formal verb ki that verb-phrasifies the preceding material Bugaeva (2012: 496); Nakagawa (2024: 374, 407). The person index belongs to the notional predicate; ki itself is unindexed in the periphrasis. The direct frame somo V places the negator before the whole person-marked verb and does not split a person prefix from the stem Satō (2008: 124); Ijäs (2023). Nakagawa finds no clear semantic difference between the two arrangements, and Satō notes that ordinary verbal negation very often uses the periphrastic pattern Nakagawa (2024: 374); Satō (2008: 22, 124).

The internal constituency of somo ki is disputed. Bugaeva and Nakagawa keep the compositional analysis in which ki is the formal verb and the preceding predicate is its complement Bugaeva (2012: 496); Nakagawa (2024: 374). Satō, in the Chitose description reported by Takahashi, treats somo ki as a single auxiliary; Takahashi leaves the historical path open Satō (2008: 86); Takahashi (2016: 74) ‹contested›. The disagreement bears on constituency while both analyses place the construction in the same post-predicate slot. The distribution and scope of the negator are developed in Chapter 126 (Standard Clausal Negation with somo).

59.3 Ki as event support and pro-verb

Outside negation, ki verbalizes event-denoting material and can stand where English translations reach for 'do'. The complement may be a predicate stem, an event nominal, or a nominalized clause, and its transitive subject is marked by the A-set, as expected for a transitive verb. This use provides the bridge between ordinary lexical ki and the negative support frame: both build a finite predicate out of material whose own syntax is already present.

The positive frame V ka ki also belongs here. Nakagawa cites ipe ka a=ki and rok=an ka ki in comparing post-predicate and preverbal negative patterns, and treats the event before ka ki as the complement verbalized by ki Nakagawa (2024: 374). These examples license a restricted pro-verb analysis: ki can carry the finite morphology for an event already specified by an adjacent complement ‹consensus›. They do not demonstrate an unrestricted substitute for a missing verb phrase; broader VP-substitution uses would require a separate corpus study ‹corpus-suggested›.

The variant iki is clearest in template diagnostics. The plural suffix -pa and the outer plural element pa form a minimal ordering pair with this verb: iki-pa=an has root-bound -pa inside the fourth-person subject suffix, while iki=an pa has the clitic after the person-inflected word Nakagawa (2024: 156–159). That ordering pair shows why support verbs still count as verbs: they take the ordinary person and number morphology when they function as the predicate of a clause, even where their lexical content is bleached.

59.4 Ki, somo ki, and clause linkage

A negated action predicate normally links to a following clause with no, yielding the common 'without doing X' reading. Nakagawa and Satō state the rule that stative predicates and negated action predicates take no, with wa excluded from this environment Nakagawa (2024: 340–341); Satō (2008: 44–47). The result is a three-level stack: ki supplies the predicate of the negated clause, somo negates it, and no links the whole to the following clause.

The same architecture appears below formal-noun nominalization and evidential predicates. A negated verb phrase can stand inside the proposition headed by the possessed formal noun ruwe; ne then predicates the nominalized proposition. The negation is internal to the nominalized clause, while the copular evidential frame scopes over it Nakagawa (2024: 393).

This ordering is a useful boundary test. In … somo ki ruwe ne the proposition is negated and the evidential remains outside the scope of negation. When somo falls on the copula of the formal-noun predicate, as in ruwe somo ne, the result is a high-scope rhetorical or convictional frame, a reading outside ordinary propositional negation Nakagawa (2024: 393). The details of that outer frame are routed to Chapter 121 (ruwe ne — the Inferential / Visual-Trace Evidential) and Chapter 126 (Standard Clausal Negation with somo). For the linking inventory see Chapter 131 (Clause Linkage in Ainu: Typology and Inventory).

59.5 Ne as copular support

The copula ne is formally transitive in valence classification, yet only its subject is indexed. Nakagawa gives okkayo ku=ne 'I am a man' with a transitive-type subject prefix and rejects object indexing on the predicate complement Nakagawa (2024: 132); Shibatani likewise describes the copula as taking transitive subject morphology Shibatani (1990: 50). Satō's three-valency analysis accounts for the pattern by separating the two-place predicate relation from indexable macroroles: the predicate nominal is a complement of the copular relation, while only the subject is available for person indexing Satō (2023: 52) ‹contested›.

Because ne is the ordinary support for predicate nominals, the negative frame of copular clauses is NP (ka) somo ne. Satō notes that ne ka somo ki, expected if the copula simply fed the ordinary somo ki frame, is almost unattested Satō (2008: 22). The copular construction keeps its own negation architecture, detailed in Chapter 96 (The Copula ne and Predicate-Nominal Clauses) and Chapter 126 (Standard Clausal Negation with somo).

The evidential example below combines both support verbs in a single sentence. Ne predicates a zero-place weather verb in the first clause, while ki hosts the support negation in the second. The layering makes it a useful diagnostic for the whole support system.

59.6 An between existential verb and support

The existential an remains an intransitive verb when it predicates existence or location, and it enters support constructions from that source. One pathway is lexical compounding into zero-place ambient predicates. Nakagawa derives mean 'be cold (weather)' from a non-free root me- plus an; such predicates take no person index because no subject slot remains Nakagawa (2024: 128–129). Shibatani analyses these as subject-incorporation, and Ijäs teaches them as avalent weather verbs Shibatani (1990: 33); Ijäs (2023). Where the first element is a free noun, the boundary is unsettled: Ijäs treats apto as 'it rains' as an avalent unit, while Nakagawa leaves the analysis open because apto is an autonomous noun ‹contested› Nakagawa (2024: 129–130).

The v0/vi pairing in this table is directly grounded in Nakagawa's analysis: the sir- prefix and the me- root are non-autonomous elements that fuse with an into a zero-place compound, whereas the felt-temperature verbs take a subject marked by the ordinary S-set Nakagawa (2024: 127) ‹consensus›.

A second pathway keeps an as an overt support verb. Existential and locational clauses are described in Chapter 97 (Existential and Locational Clauses (an / oka)); aspectual frames such as kor an and wa an are treated in Chapter 110 (The Existential-Aspect System: kor an Progressive and wa an Resultative-Perfect). In evidential constructions, the alternation between copular ne and existential an contributes to the direct/indirect opposition of formal-noun predicates, with siri an and related forms routed to the evidential chapters. These support uses preserve the intransitive origin of an while extending it into TAM and information-source constructions ‹consensus› Nakagawa (2024: 390–393, 446–449).

59.7 Ijäs and the learner-grammar claims

Ijäs's treatment foregrounds the learner's problem: Ainu negation and auxiliary-like meanings are built with independent verbs, so the person marker stays with the predicate that supplies the argument frame. Her rule that somo precedes the verb and cannot split a person marker from its stem matches Satō's and Nakagawa's description ‹consensus› Ijäs (2023); Satō (2008: 124); Nakagawa (2024: 407). Her identification of ki as a dummy 'do' verb that fills the required verb position because somo cannot stand as the predicate of a clause also matches the compositional account in Bugaeva and Nakagawa ‹consensus› Ijäs (2023); Bugaeva (2012: 496); Nakagawa (2024: 374).

Two Ijäs claims carry caveats. Her suggestion that postverbal ka somo ki reflects Japanese influence on word order is not supported by the read Hokkaido grammars and should be attributed to Ijäs alone ‹speculative› Ijäs (2023); Dal Corso (2025: 41–42). Her presentation of lexical negative verbs as blocking somo plus an affirmative counterpart also states a tendency too strongly; the primary sources describe lexicalized negative predicates and preferences, while coexistence with somo is independently attested in some environments ‹contested› Ijäs (2023).

59.8 Analytical boundaries

The support-verb system is outside the verb word proper, whose person indexes, valency prefixes, incorporated nouns, and stem-internal suffixes are ordered in Chapter 56 (The Verb Word and Its Position-Class Template). Valency-changing derivation — applicative, causative, antipassive — changes the argument frame of the lexical stem; ki, ne, and an provide a higher predicate host above the already-built stem. The lexical semantics of the copula and existential are treated in Chapter 96 (The Copula ne and Predicate-Nominal Clauses) and Chapter 97 (Existential and Locational Clauses (an / oka)). Nakagawa (2024: 451); Bugaeva (2022)

The degree of grammaticalization differs across the three verbs. Somo ki may be analysed as a compositional support-verb construction or as a single auxiliary; kor an and wa an behave aspectually while still showing clause-chain morphology; ne is a verb-like copula whose object indexing is blocked. These disagreements affect analysis, not the descriptive ordering: person-marked lexical predicate first, support predicate above or after it, and clause-level operators outside that support layer ‹contested›.