Chapter 56The Verb Word and Its Position-Class Template

56.1 The verb word as the head of its clause

The Hokkaido Ainu verb word is the head of its clause: every other constituent enters a grammatical or semantic relation with it, and it alone carries the bound indexes that identify the clause's core participants Nakagawa (2024: 124). Satō describes the language as head-marking in the sense of (Nichols 1986), with the verb bearing the central information about which arguments fill its valence frame; external noun phrases may be absent, freely ordered, or arranged for discourse, while the indexes on the verb do the ordinary work of tracking who acts on whom Satō (2023: 42). Ijäs's classroom image of verb pockets, filled before and more obligatorily than free pronouns or full noun phrases, names the same head-marking fact for learners Ijäs (2023). The personal-index architecture itself is the subject of Chapter 62 (Architecture of the Personal-Affix System: The Four Persons and the S/A/O Paradigms); this chapter sets out the verb word's morphological template and the order of its position classes, and routes the detail of each zone to the chapters that follow.

Several properties of the Ainu verb fix the boundaries of that template Nakagawa (2024: 124–125). Tense is absent as an inflectional category: a bare verb form is read as past, present, or future from context, temporal adverbs, or discourse sequencing Nakagawa (2024: 124); Shibatani (1990: 33); Ijäs (2023). There is no separate adjective class; meanings rendered by English or Japanese adjectives are intransitive verbs, and a stative reading ordinarily carries an inchoative one — Ijäs presents hure as both 'be red' and 'become red', matching Nakagawa's poro 'be big; become big' and the imperative evidence Shibatani adduces for the verbal status of quality predicates Ijäs (2023); Nakagawa (2024: 124); Shibatani (1990: 33–34). Property predicates are treated as a verb subclass in Chapter 99 (Property-Concept (Adjectival) Predication without a Copula) and Chapter 37 (The 'No Adjective Class' Thesis: Property Concepts as Stative Verbs). Person marking is obligatory while independent pronouns are optional, so the language is consistently pro-dropping and verb-final Ijäs (2023); Nakagawa (2024: 125, 363). Tense, aspect, mood, evidentiality, and negation are assembled outside the lexical verb by auxiliaries, formal nouns, and particles that take an already person- and number-inflected verb as their input, and are treated in Chapter 108 (Tense, Aspect, Mood, and Modality: Overview), Chapter 109 (Tenselessness and the Time-Reference System), and the evidential and negation parts. The verb word supplies the stem and its argument morphology; the larger predicate supplies the rest.

56.2 A position-class template

The finite verb word can be modelled as an ordered set of position classes. The model is a descriptive abstraction over the recurrent linear positions of person indexes, valency-changing morphology, the lexical stem, stem-number morphology, derivational suffixes, and the outer plural markers. Read from the outer left edge to the outer right edge, the zones are as follows. Bugaeva (2012: §15.4.4.1)

Two ordering facts anchor the template. The first is a side-switch by transitivity. The first-plural-exclusive and fourth-person subjects appear as suffixes on an intransitive stem but as prefixes on a transitive one: inkar=an 'I/one looked' and arki=as 'we came' beside a=nukar 'I/one saw it' and ci=nukar 'we saw it' Nakagawa (2024: 166–169); Satō (2008: 96). The second is the position of the two homophonous plural pa elements. The Type-C plural suffix -pa is root-bound and stands inside the person suffix and inside the causative, whereas the outer plural element pa follows the whole person-inflected verb: iki-pa=an with the suffix before =an contrasts with iki=an pa with the clitic after it Nakagawa (2024: 156–159). The relative order of these elements is the subject of Chapter 69 (The Personal-Affix Template: Position Classes and Affix Ordering) and Chapter 58 (Verbal Number: Suppletive Stems and the Pluractional -pa).

The template is partly positional and partly scope-driven. Some ordering is fixed by position class, but the valency prefixes and suffixes stack in an order that tracks the order of derivation: a chain such as ruska 'be angry at' → ko-ruska → yay-ko-ruska → i-ruska raises and lowers valence step by step, each affix scoping over the stem already built Bugaeva & Kobayashi (2022: §4); Satō (2023: 40). Auxiliaries and light verbs such as ki 'do' and the abilitative easkay 'can' stand outside this word altogether, and are treated in Chapter 59 (Light Verbs, Pro-Verbs, and the Auxiliary-Verb Construction) and Chapter 108 (Tense, Aspect, Mood, and Modality: Overview).

56.3 Transitivity as the organizing axis

Transitivity is a formal property of the verb word, decided by structure and not by the meaning of the translation Nakagawa (2024: 125–127); Satō (2008: 96). Two diagnostics define it. A verb requires a fixed number of bare noun-phrase arguments — noun phrases without a case particle or postposition — and case-marked phrases are adverbial additions outside that count; one bare argument marks an intransitive verb, two or more a transitive verb Nakagawa (2024: 126–127). An intransitive verb indexes only its subject (S), while a transitive verb indexes subject (A) and object (O) Nakagawa (2024: 126). The surest operational test puts the verb in the first-person plural: an intransitive takes the suffix =as (exclusive) or =an (fourth person), a transitive the prefix ci= or a=, the other person cells being uninformative — a side-switch Satō calls irregular and still unexplained Satō (2008: 96).

The minimal pair inkar 'look' and nukar 'see' shows why an English translation is an unreliable guide. inkar is an intransitive activity of directing the gaze with no fixed object; nukar is transitive and takes a specified object. A fourth-person subject therefore surfaces as the suffix =an after the intransitive stem and as the prefix a= before the transitive stem in one sentence Nakagawa (2024: 131):

Both arguments of a transitive verb are bare, and Ainu draws no formal distinction between a direct and an indirect object: what a Japanese translation marks as 〜を or 〜に is the same bare object in Ainu Nakagawa (2024: 127). A three-place verb such as kore 'give' takes a subject and two bare objects, all omissible, yet only two participants can be indexed on the verb — the two-index cap holds even where the clause has three arguments Nakagawa (2024: 127); Satō (2023: 39). When a non-third person is indexed it is the human recipient that the object affix tracks, a secundative pattern taken up in Chapter 66 (Object Indexing: Monotransitive and Ditransitive Clauses) Satō (2023: 41).

Ijäs's learner grammar reaches the same architecture through the contrast of verb pockets and external slots: the arguments are cross-referenced on the verb first, and a bare clause such as numan hekaci cep koyki presents both participants as bare noun phrases with an unmarked third-person verb Ijäs (2023). Her further statement that the indirect object of a three-place clause stands between subject and object is a classroom simplification; the two objects and the subject of epakasnu 'teach' occur in several orders with one valence frame and no fixed position Nakagawa (2024: 363–365).

56.4 Valence classes and what counts as a verb word

The argument count runs from zero to three, with a contested fourth. Nakagawa and Bugaeva and Kobayashi use the numeric labels 0-, 1-, 2-, and 3-argument verb precisely to keep the noun-phrase count apart from the person-affix fact of intransitive versus transitive; Satō, Tamura, and Ijäs keep the traditional intransitive/transitive/ditransitive terms Nakagawa (2024: 128); Bugaeva & Kobayashi (2022: §3); Ijäs (2023).

The zero-place class is built from non-autonomous roots: weather and ambient predicates such as mean 'be cold' (< me- 'cold' + an 'exist') and sirpirka 'be fine weather' take no participant index Nakagawa (2024: 128–129). Ijäs presents these as avalent verbs with the same examples, including the analysis of sir- 'surroundings' and a non-free me- 'cold' Ijäs (2023). Where the contributing element is itself a free noun, the analysis is genuinely open: Ijäs treats apto as 'it rains' as a single avalent unit, but Nakagawa leaves it undecided, since apto 'rain' is an autonomous noun and the string may be a two-word clause Nakagawa (2024: 128–130); Ijäs (2023) ‹contested›. The same root that gives a weather predicate often has an intransitive counterpart that does take a subject — climatic mean 'it is cold' beside experiential merayke 'feel cold' — a contrast Ijäs marks and Nakagawa lays out as a paired paradigm Nakagawa (2024: 127); Ijäs (2023).

The copula ne shows that valence is not reducible to the number of person indexes. It is two-place in the relation 'X is Y' and takes transitive-type subject marking, yet it admits no object index: okkayo ku=ne 'I am a man' is well formed while an object-indexed alternative is not Nakagawa (2024: 132); Shibatani (1990: 50). The copula and predicate nominals are developed in Chapter 96 (The Copula ne and Predicate-Nominal Clauses). Three-place verbs are mostly derived, by the applicative ko- or the causative -re, the one common base-form member being the locative o 'put X into a place' Nakagawa (2024: 134); Bugaeva & Kobayashi (2022: §4). The valence classes, the small locative-object subclass, and ambitransitivity are the matter of Chapter 57 (Transitivity, Valence Classes, and Ambitransitivity).

56.5 Valency-changing morphology in the template

The valency prefixes and the stem-internal suffixes are the positions that alter the argument count. Their syntax and semantics belong to the voice part; in the template the relevant fact is the position each occupies and the direction in which it moves the count Nakagawa (2024: 137–139); Bugaeva & Kobayashi (2022: §4).

A reflexive or reciprocal prefix occupies the object position itself, so a subject index and an object index for the same participant never co-occur Nakagawa (2024: 183), and the antipassive i- absorbs the object of a transitive stem, as in ruska 'be angry at' → i-ruska 'be angry' Bugaeva & Kobayashi (2022: §4); Nam (2021: 62–64). The prefix e- is a standing warning against reading the template off shapes alone: it can be the second-person index, an applicative, or an affiliative nominal prefix, and only the valence computation and the surrounding morphology identify it in a given word Nakagawa (2024: 138–139). The applicatives are developed in Chapter 78 (The Applicative System: Overview and the Preverb-vs-Applicative Debate), the causatives in Chapter 74 (The Morphological Causative -re / -e / -te), Chapter 76 (Causee Marking, Ditransitive and Double Causatives), and Chapter 77 (The Permissive/Sociative Causative -yar), the detransitive in Chapter 83 (The Antipassive / Detransitive i-), the reflexives and reciprocal in Chapter 84 (The Reflexive yay-), Chapter 85 (The Reflexive-Possessive si- and the yay-/si- Division), and Chapter 86 (The Reciprocal u-), and incorporation in Chapter 88 (Object (Patient) Noun Incorporation). Derivation that builds a verb stem without changing its valence — denominal and deverbal verb formation, and the adverb-forming -no that turns a property verb into an adverb (poro → poronno) — is treated in Chapter 60 (Denominal and Deverbal Verb Derivation) and Chapter 61 (The Adverbializer -no and Adverb Formation from Verbs).

56.6 The stem zone and verbal number

The stem zone holds the lexical root together with any suppletive or suffixal singular/plural alternation. Number in Hokkaido Ainu sits on the verb, not the noun: nouns are number-neutral and read as singular or plural by context — Ijäs notes that cep 'fish' may be either — while a minority of verbs have distinct singular and plural stems Nakagawa (2024: 124–125, 140–142); Ijäs (2023). The alternation takes three shapes: outright suppletion (ek / arki 'come', an / oka 'exist'), the -n / -p pattern of directed-motion verbs (san / sap 'go down'), and the vowel-to--pa pattern of change-of-state transitives (tuye / tuypa 'cut') Nakagawa (2024: 146–150); Satō (2008: 99–106).

The two transitivity classes index plurality differently. For intransitives the alternation is close to agreement with S: a plural subject normally selects the plural stem and a singular subject the singular stem Nakagawa (2024: 145). For transitives the plural stem marks a plural object or a pluralized or repeated event, with the singular stem remaining the default even where the object is plural; Shibatani, following Durie, treats the alternation as a selectional restriction on the principally affected argument — S of an intransitive, O of a transitive — rather than as agreement Nakagawa (2024: 142–145); Satō (2008: 101); Shibatani (1990: 51–53). Numerals and quantifiers show why the slot resists a simple agreement rule: a subject numerically specified as two, three, or sometimes four ordinarily takes a singular intransitive stem, against counterpatterns with iwan 'six' and with poronno 'many' that vary by dialect and speaker Nakagawa (2024: 160–162) ‹contested›. The full inventory, the corpus frequencies, and the agreement-versus-selection debate are the subject of Chapter 58 (Verbal Number: Suppletive Stems and the Pluractional -pa), and the interaction with person of Chapter 70 (Person–Number Interactions and Plural Agreement).

56.7 Debates that shape the template

Four analytical questions bear on how the template is read, and none is settled. The first is the mechanism behind the valence calculus. Nakagawa's arity assigns every element — base verb, each derivational affix, each person index — a positive, negative, or zero value, the sum predicting the verb's remaining bare-noun-phrase positions Nakagawa (2024: 136–139). Satō argues that a single arity value fails for the copula, for auxiliaries, and for ditransitives, and replaces it with three simultaneous layers — an invariant argument-slot count, a macrorole valency in the sense of (Foley & Van Valin 1984) that the person affixes index, and the morphological valency of the derived word — while flagging his own scheme as something still to be justified Satō (2023: 42–52); Satō (2025: 22–26) ‹contested›. The template here records position and valence effect without committing to either mechanism.

The second question is the third-person zero. Nakagawa assigns the third person no overt affix and no arity value, so that the calculus balances; Satō holds that a zero third-person affix is needed to explain the two-index cap and the impossibility of certain three-participant combinations Nakagawa (2024: 137); Satō (2023: 39); Satō (2025: 23–24) ‹contested›. The transitivity table above leaves the third-person cells empty and routes the analysis to Chapter 62 (Architecture of the Personal-Affix System: The Four Persons and the S/A/O Paradigms).

The third question is the outer plural pa. Nakagawa and the Handbook treat the postposed pa, eastern ci, and the Sakhalin cognate as a single clitic class distinct from the root-bound suffix; Satō's Chitose data make the clitic analysis doubtful, since postposed -an pa is rare and a speaker rejects *yaynu-as pa in favour of yaynupa-as Nakagawa (2024: 156–159); Satō (2025: 25–26) ‹contested›. The fourth question is the four-place class: Nakagawa lists causativized ditransitives such as omarere as four-place verbs, while Satō warns that the type is too marginal to support a regular category Nakagawa (2024: 135); Satō (2025: 22–23) ‹contested›.

These debates leave the linear template stable while its interpretation is open. Ainu has very few bare stems that serve as both intransitive and transitive without overt marking, with roski 'stand' effectively the sole instance, so valence alternation is almost always coded by the derivational positions catalogued above rather than by lability Satō (2008: 97); Bugaeva & Kobayashi (2022: §5). The valence classes and ambitransitivity are pursued in Chapter 57 (Transitivity, Valence Classes, and Ambitransitivity), and the productive derivation in Chapter 73 (Valency, Voice, and Noun Incorporation: Overview).