Chapter 17Negation

17.1 The negators and their composition

The negative forms attested in the corpus of Sakhalin Ainu are listed below with the morphological analyses of Dal Corso (2025b: 46–47). All of them contain the element ham, which Dal Corso analyses as a proclitic ham= on the grounds that it is not class-selective — it attaches to verbs, nouns, the copula, adverbs, and particles — and that it is phonologically dependent on its host.

The proclitic shows the regular nasal assimilation described in Chapter 5: /m/ assimilates in place to a following consonant, so ham=ne yields hanne and ham=ki yields hanki, while before s it dissimilates to y, ham=suy ‘not again’ yielding haysuy — the same alternations the nasal-final clitic an= shows (Dal Corso 2025b: 45–46). Two further observations bear on the clitic analysis. The first is prosodic: ham= ordinarily fuses with its host into one phonological word, re-syllabifying it and drawing the stress onto itself (hánne, háysuy); but Piłsudski’s accented spelling hám utara for ham=utara ‘not the people’ points to a pronunciation [ˈha.mu.ta.ɾa] where the regular stress rules of a fused CV.CV word would give *[ha.ˈmu.ta.ɾa] — a hint that the negator could still carry stress of its own, as befits an erstwhile independent word recently reduced to clisis (Dal Corso 2025b: 45–46). The second is positional: on verbs, ham= stands outside the person prefixes — Piłsudski records hamecinu ‘you did not listen’, that is ham=eci-nu, and hamanrayki ‘I did not kill’, that is ham=an-rayki — a position at odds with ordinary verbal prefixes such as the applicative or the reflexive, which follow agreement. Only two tokens in the whole record show the reverse order person–ham: anameyaynu (an-ham-e-yay-nu) ‘I did not pay attention to it’, from Takoye (Majewicz 1998: 278), and ’anhancinukahsiri ‘a land I did not know’, from Rayciska. Dal Corso leaves open whether these show ham= on its way to becoming a true inflectional prefix or merely a negator frozen inside a lexicalized word form, on the model of Hokkaidō somoipere ‘not feed’ (Dal Corso 2025b: 46).

17.2 Standard negation

Direct procliticization as in (1) is essentially an East-coast pattern: 19 of its 26 verbal tokens come from East-coast texts, and in the West-coast corpus it survives only inside the folklore formula hancinukahsiri ‘an unseen land’ (Dal Corso 2025b: 49–50). Everywhere else, declarative negation is analytic. The dominant construction — about a third of all negative tokens — places the negated verb first, usually followed by ka ‘even’ or its emphatic variant kayki, and closes the clause with hanki, the negated light verb ‘do’ (Chapter 12 on ki ‘do’). The notional verb is a zero-nominalized object of ki, and the two coasts disagree over where person marking goes. In the older layout, agreement sits on ki, the syntactic head, and the notional verb stays bare; in the layout that dominates the West-coast corpus, agreement has migrated onto the nominalized notional verb, leaving ki uninflected (Dal Corso 2025b: 50). The next two examples show the contrast:

Dal Corso reads the westward migration of agreement onto the notional verb as a sign that the construction was being reanalysed from two predicates into a single predicate with a postverbal negator (Dal Corso 2021); (Dal Corso 2025b: 57–58); agreement marked on both verbs at once occurs exactly once in the corpus (Dal Corso 2025b: 50).

The same light-verb frame admits the univerbated negators of the hanne family in place of hanki (Dal Corso 2025b: 51–52):

The same analytic logic appears in the Dobrotvorsky-era texts of the 1860s–70s, the oldest connected West-coast record, where ham negates the light verb in the intentional construction kun(i) ki kusu (Chapter 15):

Alongside the light-verb pattern, the univerbated forms hanne (East coast), hanneh, and hannehka (West coast) are used preverbally, immediately before the negated verb:

The corpus offers no evidence for a dedicated interrogative negation: the strategies above appear unchanged in questions (Dal Corso 2025b: 49). Negative indefinite readings — ‘not a single, none at all’ — are built with a numeral or interrogative plus ka(yki) ‘even’ under a negated predicate (Chapter 24 on ka):

17.3 Negation of the copula

Equative clauses (Chapter 18) are negated on the East coast by procliticizing ham= to the copula ne, yielding hanne:

In the West-coast recordings this is impossible — ham= is no longer productive there — and the grammaticalized form hannehka has taken over the function of a negative copula:

17.4 Negative existence and the lexical negatives

Existence is not negated with ham=. Sakhalin Ainu instead has a closed set of lexically negative stative verbs, each paired with an affirmative verb that is not normally negated analytically (Dal Corso 2025b: 40):

The negative existential isam (East-coast spelling often isan) covers ‘there is no X’, ‘X is absent, away’, and — with a possessor — ‘X has no Y’:

Following another verb linked with wa, isam has grammaticalized into a conclusive aspect marker, ‘do completely, off, away’ (Dal Corso (2025b: 41); Chapter 15). The pattern is extremely common in Asai Take's narratives:

17.5 Prohibitives, apprehensives, and irrealis negation

The prohibitive is formed by placing hanka (West coast) or hankayki (East coast) before a verb unmarked for person, exactly like the affirmative imperative (Chapter 15); the command may be reinforced by a final particle such as wa(a), kanne, or yan (Dal Corso 2025b: 53–54). Sakhalin thus builds its prohibitive from the ordinary negative element, whereas the Hokkaidō dialects use the dedicated prohibitive word iteki and its variants (Dal Corso 2025b: 54). The source of the Sakhalin marker is typologically remarkable: prohibitives are usually recruited from verbs of abstaining or stopping, from modals and desire predicates, or from imperative morphology, whereas hanka(yki) is the negation of the additive particle ka(yki) ‘even’, as the table of forms above records, a path with only scattered parallels elsewhere and one that Dal Corso traces through the negative contrastive-focus use of negated particles (Dal Corso 2025b: 60). The segmentation itself is old: Chiri already parsed the form as negative ham plus emphatic ka (Chiri 1942: 536).

In the West-coast corpus hanka is not confined to commands addressed to an unmarked verb. The corpus shows it with person-marked verbs in directive contexts, where the negated clause functions adverbially; we take this as a bridge between prohibition and general irrealis negation:

The East-coast dialect of Hunup places hanna (variant hana) before a verb to express apprehension — ‘lest…’ — and, when the verb carries the intentional auxiliary kusu iki, negative intention (Dal Corso 2025b: 54–55):

The same dialect uses hannah to negate the condition of a conditional sentence. The protasis closes with the causal-final linker kusu and the topic particle neyke; hannah follows, before a second kusu that partially restates the conditional nexus — in effect the negative-focus syntax of the next section turned on a whole clause. Where the apodosis predicate is nominalized and bears the conditional suffix -kun, as below, the sentence is counterfactual: the outcome failed to come about because the condition never held (Dal Corso 2025b: 55–56).

17.6 Caritive and focus negation

The form hamo ~ hamu — historically ham plus the adverbializer -no — negates a bare, non-finite verb that stands in an adverbial relation to the following predicate, ‘without doing’ (Murasaki (1979: 116); (Dal Corso 2025b: 56)). No clause linker appears, and the negated verb carries no person marking; the caritive form itself signals the dependency (Chapter 23).

No split form *ham no is attested, so the segmentation is a reconstruction; it is nonetheless well supported. The -no is the ordinary manner adverbializer, and the missing nasal follows from the same lenition seen in ponno ~ pono ‘a little’; the Hokkaidō negator somo continues the same Proto-Ainu verb through just such an adverbialized form, *sam-no, whose final stress explains its rounded vowel; and hamo itself keeps first-syllable stress despite its CV.CV shape, the unstressed final vowel rising to give the variant hamu — the profile of a stressed negator univerbated with a stressless suffix (Dal Corso 2025b: 44–45, 47). Chiri supplies an inner-Ainu parallel: in Tokachi (Hokkaidō), sem-ki-no ‘without doing’ contracts to sinkino, the s-form negator feeding a caritive adverbial in -no by the same route (Chiri 1942: 536).

Finally, the East-coast texts attest ham= directly on a noun — always utara ‘people’ — on the adverbs suy ‘again’ and ene ‘like this’, and once on the topic particle neampe, marking contrastive focus negation: the marked constituent is singled out as not satisfying the predicate, against some alternative that does — ‘not X (but Y)’ (Dal Corso 2025b: 52–53):

17.7 History of the system

Dal Corso reconstructs the whole family of negators from a Proto-Ainu stative transitive verb *sam ‘not exist in (a place)’ (Dal Corso 2025b: 43–45). One line of development antipassivized the verb with i- (Chapter 14), yielding the negative existential isam ‘not exist (anywhere)’; the other turned the bare verb into a negative marker placed before the negated element, which weakened to the proclitic ham=, with regular *s > h outside the neighbourhood of i. The Hokkaidō negator somo continues the same verb through an adverbialized form *sam-no; the Sakhalin caritive hamo is its direct cognate, as set out in the previous section.

Dal Corso sets this prehistory within two named typological cycles. The step from negative verb to negative marker instantiates the Givón Cycle, by which erstwhile negative verbs are worn down into negators (Givón 1978). On Croft’s Negative Existential Cycle (Croft 1991), the reconstructed stage in which one form, *sam, served as both verbal negator and negative existential is Type C — the rarest of Croft’s types, reported otherwise mainly from Dravidian and Polynesian languages (Veselinova 2016) (Dal Corso 2025b: 56–57). The split into existential isam and verbal ham=, driven by the antipassive and by the sound change, took Sakhalin Ainu directly to Type B, with distinct negators for the two domains, skipping the intermediate Type A in which one strategy covers both; since the stages of the cycle need not be sequential, the skip is unexceptional (Dal Corso 2025b: 57).

From procliticized ham= (Stage 1) the analytic system then unfolded in implicational order (Dal Corso 2025b: 57–59). Negation of the copula gave hanne, which in the East was reanalysed as a preverbal verbal negator and drawn into the light-verb frame to mark emphasis (Stage 2); because hanne was still felt to be copular, the construction was renewed with overt nominalization as hanneh (ham=ne-p) and emphatic hannehka (Stage 3). In the West, where ham= had ceased to be productive, preverbal hanneh(ka) was then pressed into service as the plain, non-emphatic negator (Stage 4) — an emphatic form bleaching into the default, the profile familiar from the Jespersen cycle of negator renewal. Finally, by ellipsis of ki, hannehka came to stand as a negative copula in its own right (Stage 5): a verbal negator turning copular with no new affirmative existential beside it, a backward transition in the Negative Existential Cycle that Dal Corso suggests is so far unparalleled (Dal Corso 2025b: 58–59). The light-verb frame itself need not be a purely internal invention: negation through a nominalized predicate governed by a dummy verb recurs in Hokkaidō Ainu, in Nivkh, and in Chukotko-Kamchatkan, and diffusion around the southern Okhotsk Sea is a live possibility (Dal Corso 2021: 223–225); (Dal Corso 2025b: 57–58).

The corpus counts lay the cline out on the ground. Every later stage is the stronger represented the further west and the later the recording: synthetic ham= and hanne are essentially Eastern, hanki and hannehka overwhelmingly or exclusively Western (Dal Corso 2025b: 47).

The same drift can be read against time within the western records themselves. Of the negation tokens in Dobrotvorsky’s West-coast materials of the 1860s–70s, the preverbal clitic strategy still accounts for 68.4%; in Murasaki’s 1983 data it has fallen to 3.7%, while the postverbal light-verb construction has risen from zero to 49% (Dal Corso 2021: 223). Taken at face value, the figures compress the synthetic-to-analytic turnover into roughly forty years, which Dal Corso himself flags as suspiciously fast; more plausibly the two dialect groups moved at different speeds, the East conservative and the West innovating, so that the staged cline is as much a geographic gradient as a chronological sequence — an implicational ordering of overlapping strategies rather than a dated relay (Dal Corso 2025b: 61).

The Hokkaidō negator somo itself surfaces in Sakhalin sources only marginally: Dal Corso finds it in Piłsudski's Tunayci texts, told by speakers who had lived in Hokkaidō, and treats it as contact influence (Dal Corso 2025b: 41). The corpus adds one more such setting: in Sentoku Tarōji's letters of 1906 the fixed conditional phrase somo ne ciki ‘if (it) is not (so)’ appears beside native ham-negation, fitting the same profile of a literate East-coast writer in close contact with Hokkaidō usage.