Chapter 20Nominalization

20.1 Nominalizing words

The bound nouns pe ~ p ‘thing, one’ and hi ~ i ‘thing, place, time’ head participant and event nominalizations. With pe the result ranges from fully lexicalized nouns to transparent headless relatives ‘the one that…’ (Chapter 21):

Event nominalizations in hi as in (2) feed topicalization, complementation, and the conditional-irrealis form kun(i) pe (Chapter 22). The same bound nouns are fossilized inside core grammar: the West-coast negator hanneh(ka) contains ham=ne-p ‘NEG=COP-thing’ (§17.1), and the topic particle neampe is a frozen ne-an-pe ‘the thing which is’ (Chapter 24). The place noun usi(ke) ‘place’ supplies locational nominalizations, surviving in lexicalized place-words; productive examples in the Sakhalin record are scarce, and locational nominalization is normally carried by hi or by relativization on a locative noun (Chapter 21).

20.2 The possessive nominalization

Sakhalin Ainu extended the possessive form beyond the noun. Where Hokkaidō varieties possess only a subset of nouns, Sakhalin generalized the suffix to all nouns — an=mosir-ihi ‘our country’ (Murasaki (1979: 5); (Bugaeva 2016: 105)) — and then carried the same morphology onto verbs, yielding a genuinely non-finite form V-POSS that has no counterpart elsewhere in Ainu (Bugaeva 2016: 104–105). The suffix copies the final vowel of the stem with epenthetic h, exactly as on nouns (§5.2). Murasaki had recorded these verb forms as a separate particle hVV of assertion and sympathetic questioning (Murasaki 1979: 72); Bugaeva identified them with the possessive paradigm and unified their uses (Bugaeva 2016: 112). We adopt her analysis, and gloss the form POSS throughout the grammar.

As a clausal argument, the nominalized verb stands where a noun phrase would:

20.3 Adverbial and modal hosts

Most occurrences of the possessive nominalization are governed by a small set of conjunctions and auxiliaries that contain the copula ne(e): manner neeno ‘as, like’, adversative neya ‘but’, concessive neyah ‘although’, affirmative neeko ‘surely’, dubitative nean and nee nankor ‘probably’ (Bugaeva 2016: 106–107). Their Hokkaidō cognates, where they exist, govern finite verbs; the Sakhalin pattern V-POSS + ne-form is the innovation.

20.4 Nominalization in the evidential construction

The evidential sentences of Chapter 16 are themselves built on nominalization: the possessed evidential nouns ruwehe, hawehe, sirihi, humihi head the clause, taking the scope proposition as their possessor and existential an as predicate, on the template [possessor clause]–N-POSS–an. In the basic pattern the possessor is a plain finite clause and no overt nominaliser appears on the scope verb at all — re monimahpo an ruwehe an ‘evidently three girls lived there’, literally ‘the trace of three girls being there existed’ (§16.3). Sakhalin diverges from Hokkaidō in also admitting a copula-linked subtype, in which the lexical verb is overtly nominalised with the possessive suffix and joined to the evidential noun by the copula nee — V-POSS nee N-POSS an (Bugaeva 2016: 108):

20.5 Main-clause nominalizations

The possessive nominalization also stands sentence-finally with no copula and full illocutionary force — an insubordination pattern. Bugaeva's survey of conversation finds it in assertions, polar questions, content questions, and exclamations, and identifies the common thread as presupposition: the nominalized form presents its content as established or taken for granted, while finite forms carry new information (Bugaeva 2016: 109–112):

Furukawa's Rayciska data show the same form under her ‘past’ reading — tara aynu omanihi ‘that person went / did he go?’ — and the question particle aa selects precisely this nominalized form (Furukawa (1967: 109–110); §19.4). The perfect-like flavour follows from the presuppositional semantics: an event presented as settled fact.

That the sentence-final form really is a nominal, and not a finite verb that happens to resemble one, can be shown rather than assumed. The interrogative final particle hetaneya attaches only to nouns; yet it accepts the nominalized verb as its host, as in the rhetorical question an-i-hi hetaneya ‘why on earth is there…?’ of §16.6. Dal Corso reads this as the decisive diagnostic and analyses the sentence-final V-POSS (and V-POSS + copula) patterns as true insubordination: formally subordinate clauses conventionalized as independent assertions of personal knowledge (Dal Corso 2018: 134–135).

The main-clause use is also where the two coasts part ways. The Rayciska record shows the full range of the table in the chapter introduction, insubordinated assertions and questions included. The East-coast texts deploy the nominalization freely in its embedded functions — as a clausal adjunct under causal kusu, e.g. naha nuu-hu kusu ‘because of his hearing that’, and as the object of a cognition verb, e.g. the …-pi-hi nominalization embedded under an-eramuskari ‘I do not know’ (Piłsudski 1912) (Chapter 22) — but a bare sentence-final assertion or question of the West-coast type is not clearly attested in them. We read the asymmetry diachronically: the embedded uses belong to the older stages of the grammaticalization path traced in §20.6, the insubordinated main clause to the youngest, so the less innovative East retains the construction precisely where it arose first. A securely main-clause East-coast token would overturn this reading; we have not found one, and we offer the gap as a provisional dialectal finding rather than an accident of the record.

20.6 History

Bugaeva orders the Ainu nominalization strategies into a grammaticalization sequence: zero-nominalization of the bare verb is the oldest layer; the nominalizing-word strategy of Hokkaidō is the second; and the Sakhalin possessive nominalization, first in copula-based adverbial clauses, then in the evidential construction, and finally as a main-clause form, is the youngest (Bugaeva 2016: 113). All three layers coexist in Sakhalin Ainu, which thus preserves the history of the category in its synchrony. She raises the possibility that the non-finite verb form arose under the influence of the Tungusic languages of Sakhalin, with their fully developed non-finite subordination (Bugaeva 2016: 114); the contact setting is real (§1.4), and we consider the hypothesis plausible but not yet demonstrated — settling it would require dating the rise of the verb-possessive forms against the chronology of Ainu–Uilta bilingualism.