Chapter 27Phrasal and Utterance Intonation

27.1 Rising-nucleus accent in utterance context

Hokkaido Ainu has a pitch-accent system in which the phonologically distinctive unit is the syllable at which F0 rises — a rising-nucleus (上り核アクセント) system — in contrast with Standard Japanese, whose falling-nucleus system marks the syllable at which F0 subsequently drops Nakagawa (2024: 48–49); Shiraishi (2022: §5). The word-level placement rules for the accent nucleus are stated in Chapter 23 (The Pitch-Accent Placement Rule and the Accented/Accentless Dialect Split); the theoretical question of whether the system is best analysed as culminative accent or as an autosegmental tonal melody is addressed in Chapter 24 (The Prosodic Unit and the Accent-versus-Tone Analysis).

The rising-nucleus property means that F0 may remain high or drop after the accented syllable without changing the lexical identity of a word — only the position of the rise is phonologically distinctive. For the Saru minimal pair nísap 'suddenly' vs nisáp 'shin', nísap may be realized with sustained high pitch on both syllables (nísáp) without shifting to nisáp; only a realization in which the lowest F0 point falls at σ1 would misidentify the word Nakagawa (2024: 48–49). Community discussion illustrates the same point with the pronoun ku=ani 'I', whose accent falls on á (σ2, by regular rule): ku↗ani and ku↗a↘ni are both acceptable; ku↘ani, with the F0 minimum at the accented syllable, is not aynu-corpora Discord (2023–2026) (nukopoli, 2024-12-09).

(Nakagawa (2024: 51); Shiraishi (2022: §5.1).) Words in the exception class — which includes nísap, nína, húre 'red', sísam 'Japanese person', and a score of others — have σ1 open (CV) yet accent σ1 instead of the regular σ2. Their Sakhalin cognates show a corresponding long vowel at σ1: ‹SA› niisah vs nisah, huure, siisam (Nakagawa 2024: 27, 51). The diachronic relationship between Hokkaido accent and Sakhalin vowel length is discussed in Chapter 160 (Hokkaido Dialect Microvariation: Phonology and Morphosyntax); additional minimal pairs are catalogued in Chapter 25 (Lexical/Contrastive Accent and Minimal Pairs).

The accented syllable is slightly lengthened in polysyllabic words, more markedly in north-eastern dialects (Asai 1972/1976, reported via Shiraishi (2022: §5)). This allophonic lengthening explains older Japanese-script representations such as ユーカラ for yukar, where the long vowel corresponds to the accent on σ1 Nakagawa (2024: 26).

27.2 The accent-dialect split

Lexically contrastive pitch is found only in the accent dialects. Nakagawa and Fukazawa Nakagawa & Fukazawa (2022: §3.1) distinguish three prosodic classes: the standard-accent dialects (Saru, Chitose, Horobetsu, Asahikawa, Nayoro, Obihiro, Sōya), the irregular-accent dialects (Yakumo, Oshamambe, where the accent shifts one syllable rightward when both σ1 and σ2 are open), and the unaccented (無アクセント) varieties (Shizunai, Urakawa, Samani, Bihoro, Kushiro, Shiranuka), in which pitch is not lexically distinctive Nakagawa (2024: 49–50). The dialect classification is treated in Chapter 159 (Hokkaido Dialect Classification and Dialectometry). The utterance-intonation descriptions below rest on data from Saru and, to a lesser extent, Chitose; their applicability to the unaccented varieties is undocumented.

27.3 Declarative and continuation contours

In Saru declarative utterances, F0 follows a natural declining course from the beginning to the end of the utterance, without a sentence-final intonation gesture — neither a terminal rise nor a marked fall (Tamura 1996: 4, reported via Shiraishi (2022: §7)). Acoustic measurements on Saru speech confirm this pattern (Tamura and Motohashi 2002b; Makarova 2002, both reported via Shiraishi (2022: §7)). Within this declination, the successive rising-nucleus accents of individual words produce a series of local F0 elevations against a gradually lowering baseline.

Simeon (1968, reported via Shiraishi (2022: §7)) characterizes the prosodic status of non-final clause boundaries in Ainu: a level or falling contour signals that the utterance is not yet complete and more follows, while a rising terminal marks the utterance as prosodically finished. Under this description the default Saru declarative — ending with gradual declination but no added terminal gesture — has the profile of a continuation contour, and a clause that ends with a terminal rise (as in a polar interrogative or an exclamatory utterance) is the marked, utterance-final form.

27.4 Interrogative contours and sentence-final particles

Polar interrogatives in Saru carry a rising final intonation, with the F0 and intensity peak localized on the sentence-final particle. An acoustic study of 35 Saru interrogative utterances found that the peak falls on the last word, normally one of the interrogative particles hawe, ya, a, or he (Makarova 2002, reported via Shiraishi (2022: §7)). Tamura's Saru description extends the rising pattern to any utterance in which the speaker expects a reaction — a question, a report from the listener, a request, or an invitation (Tamura 1996: 4, reported via Shiraishi (2022: §7)). The sentence-final particles themselves are described in Chapter 143 (The Sentence-Final Particle System and Illocutionary Force); interrogative strategies and the question particles are treated in Chapter 144 (Interrogative Strategies, Question Particles, and Evidential Questions).

One exception: the particle ya appears to select a falling rather than a rising contour in some Saru data. A form transcribed as amkir_a (from amkir ya 'have you been?') was realized with falling pitch at the utterance end (Tamura and Motohashi 2002b, reported via Shiraishi (2022: §7)). A contact explanation is plausible: Hokkaido Japanese uses falling intonation even in yes/no questions, and this influence has been suggested for some falling interrogative realizations in Ainu Shiraishi (2022: §7) ‹contested›. Whether the falling realization of ya-questions is a property of the particle, a speaker variable, or a contact phenomenon is not settled on current evidence.

A further contour type is documented when the sentence ends in a nominalized or conjunctive form without a final predicate verb. In such cases the final F0 is level or falling, and the utterance communicates incompleteness or a request for confirmation rather than a settled assertion (Shiraishi (2022: §7), drawing on Tokachi data). This aligns with the Simeon generalization noted above: absence of a terminal rise leaves the utterance prosodically open.

27.5 Exclamatory and wish prosody

Exclamatory and optative utterances in Saru are produced at a pitch level elevated above the normal declarative baseline throughout, with the final syllable or syllables carrying a lengthened rise-fall (Tamura 1996: 4, reported via Shiraishi (2022: §7)). The elevation is applied to the whole utterance, not only to the final position, so it interacts with the lexical accents of each word without displacing them. The following Chitose example, drawn from Nakagawa's grammar, illustrates the construction:

The same elevated-register pattern extends to optative and invitation utterances, which share the whole-utterance F0 raise without necessarily carrying the lengthened final gesture.

27.6 Phrasal grouping and prosodic domains

No published instrumental study addresses intermediate prosodic phrasing in Hokkaido Ainu — the question of whether there are prosodic phrase boundaries between the word and the full utterance at which F0 resets, phrase-final lengthening occurs, or pauses are inserted systematically. The available descriptions treat the utterance as the domain of F0 declination and the prosodic word as the domain of accent.

Two word-level prosodic facts bear on this question. First, the person-affix paradigm distinguishes clitics by their degree of prosodic integration with the following stem. Affixes ku=, e=, and ci= form a single prosodic word with their host and shift the stem's accent; a= and eci= do not shift accent and are prosodically less integrated (Kindaichi 1931; Tamura 1970, both reported via Shiraishi (2022: §5.1); Nakagawa (2024: 52–53)). This clitic-class distinction is treated in Chapter 62 (Architecture of the Personal-Affix System: The Four Persons and the S/A/O Paradigms) and Chapter 26 (Accent in Compounds and under Affixation and Cliticization); it constitutes a gradient of prosodic bonding that may correlate with phrase-boundary placement in connected speech.

Second, personal suffixes such as =as (exclusive 1pl) and =an (4th person S) carry a secondary accent alongside the primary stem accent in open-syllable stems: iˈruska=ˌas 'we (excl.) are angry', ipˈe=ˌan (Kindaichi 1931; Tamura 1970, reported via Shiraishi (2022: §5.1)). The secondary accent is lost when the stem ends in a closed syllable: ˈcis=an 'we cry', ˈsap=an. The prosodic word template and the suffix interactions are treated in Chapter 56 (The Verb Word and Its Position-Class Template).

27.7 Typological comparison: Hachijō intonation

Community discussion of Ainu prosody has noted that the intonation patterns of Hachijō-shima Japanese feel "very close to Ainu" (aynu-corpora Discord (2023–2026) nukopoli, 2024-12-08). Hachijō Japanese is known from the Japanese dialectological literature to have a rising-nucleus pitch-accent system, making it structurally parallel to the Hokkaido Ainu standard-accent dialects and distinct from the falling-nucleus system of Standard Japanese. The perceived intonational resemblance is therefore typologically plausible — both languages share the consequence that the declarative default involves gradual F0 decline without a marked terminal gesture, and that interrogative intonation involves a terminal rise — but no published comparative acoustic study of Hachijō and Ainu intonation has been consulted for this grammar. The observation is reported here at community grade: asserted and typologically grounded, but awaiting instrumental verification ‹corpus-suggested›.