Chapter 23The Pitch-Accent Placement Rule and the Accented/Accentless Dialect Split

23.1 The rising-nucleus system

Hokkaido Ainu has a pitch-accent system in which the prominent syllable is marked by a pitch rise, not a pitch fall. The theoretical distinction matters because Standard Japanese pitch accent is analysed as a falling nucleus (下り核) — the distinctive cue is where pitch drops — whereas Ainu has a rising nucleus (上り核アクセント) — the distinctive cue is where pitch rises from low to high Nakagawa (2024: 48–49); Shiraishi (2022: §5). A word's accent identity is therefore fixed by the syllable at which the L→H transition occurs; post-nuclear syllables may continue high, or fall at the utterance edge, without changing the word's accent class.

The contrast is illustrated by the near-minimal pair nísap 'suddenly' (accent on σ1) vs nisáp 'shin' (accent on σ2). nísap can be produced with both syllables at high pitch (nísáp) without changing its identity, because the rising-nucleus cue is already satisfied at σ1; the same flat-high production of nisáp would be heard as 'suddenly', since the σ2-rise is missing Nakagawa (2024: 48–49). The Ainu accent system thus records where pitch begins to be high; the Japanese system records where it ceases to be high.

Accent is marked in this grammar with an acute diacritic on the vowel of the prominent syllable: kamúy (accent on σ2), áynu (accent on σ1). The same event is representable with a rise arrow: ka↗muy, ↗áynu. In IPA, a high-tone mark (◌́) on the nucleus of the accented syllable expresses the same. Community participants summarize the typological point concisely: ku↗a–ni and ku↗a↘ni are equivalent pronunciations of the first-person pronoun; ku↘ani is deviant aynu-corpora Discord (2023–2026) (nukopoli, 2024-12-09; corroborated by Nakagawa (2024: 48–49)). Accent marks appear in citation forms and interlinear examples throughout this grammar and are suppressed in running text.

The theoretical question of whether the system is better modelled as a culminative accent (one H per prosodic word) or as an autosegmental tonal melody (L…LH…) is examined in Chapter 24 (The Prosodic Unit and the Accent-versus-Tone Analysis). Intonation and sentence-level prosody are treated in Chapter 27 (Phrasal and Utterance Intonation).

23.2 Dialect classes

The pitch-accent placement rules described in this chapter apply to the standard-accent dialects. Hokkaido Ainu divides into three prosodic classes Nakagawa & Fukazawa (2022: §3.1); Nakagawa (2024: 49–50):

Nakagawa (2024: 49–50) includes Bihoro, Shizunai, and the eastern Hidaka localities (Samani, Urakawa) among the accentless varieties. Refsing (1986: 73) had earlier classified Shizunai as an accent dialect with rules analogous to Saru; archival evidence — scattered accent marks from a single Shizunai speaker showing both nisáppone and nísappone for 'suddenly' — supports Nakagawa's accentless classification Nakagawa (2024: §1.4.2) ‹contested›. Unaccented varieties and the phonological consequences of accent loss are treated in Chapter 160 (Hokkaido Dialect Microvariation: Phonology and Morphosyntax).

23.3 The placement rule

In standard-accent dialects, the position of the accent depends entirely on the structure of σ1 — the first syllable of the prosodic word — regardless of total word length Nakagawa (2024: 50); Shiraishi (2022: §5.1); Satō (2008: §2.6). Two rules cover the system:

Rule 1: If σ1 is closed — i.e. has the shape (C)VC — σ1 bears the accent.

Rule 2: If σ1 is open — i.e. has the shape (C)V — σ2 bears the accent.

Shiraishi (2022: §5.1) recasts these as an iambic (LH) pattern: a light (open) first syllable predicts accent on σ2; a heavy (closed) first syllable predicts accent on σ1. Kindaichi and Chiri reached the same characterization: "the iambic rule is dominant" (Kindaichi & Chiri 1936: 11, reported via Shiraishi (2022: §5.1)).

The rule is strictly local to σ1: no further syllable weight is consulted. A word of any length whose first syllable is closed carries σ1-accent (yáykosiramsuypa, nine syllables but σ1 = yay [CVC] → σ1-accent); an open σ1 gives σ2-accent regardless of what follows Nakagawa (2024: 50).

In Yakumo, Rule 2 applies one syllable further rightward when both σ1 and σ2 are open: standard notákam corresponds to Yakumo notakám; standard sinúye to Yakumo sinuyé; standard hoyúpu to Yakumo hoyupú. When σ2 is closed, Yakumo and standard dialects agree (parúnpe in both). Some Yakumo forms are additionally exceptional (hopúni, osórusi) Nakagawa (2024: 50–51); Nakagawa & Fukazawa (2022: §3.1).

Accent placement in compounds and under person-marking prefixes, where the rules reapply over the enlarged prosodic word, is covered in Chapter 26 (Accent in Compounds and under Affixation and Cliticization). The syllable shapes that feed these rules — closed vs open, and the consonant-status of coda -y and -w — are established in Chapter 21 (Syllable Structure, Phonotactics, and Word-Edge Constraints).

23.4 The pitch contour

The phonetic realization of the accent in standard-accent dialects follows a consistent contour: syllables preceding the accent syllable are at low pitch; from the accent syllable onward pitch is high. In connected speech a fall at the utterance boundary is typical, but this is a phrasal boundary effect and is not the cue that identifies the accent class. What is contrastive is solely the location of the L→H transition Nakagawa (2024: 48–49); Shiraishi (2022: §5).

For a two-syllable word with σ2-accent, such as kamúy [ka(L) ↗muy(H)], the first syllable is produced at a low pitch and the rise occurs at the onset of σ2. For a word with σ1-accent, such as áynu [↗áy(H) nu(H or ↘)], the pitch is high from the start of the word. The community characterization of the overall pattern — syllables before the nucleus are low; from the nucleus onward pitch stays high; a fall occurs toward the end of the prosodic phrase — contrasts with the Tokyo Japanese pattern in which the diagnostic event is a drop after the nucleus aynu-corpora Discord (2023–2026) (nukopoli, 2024-12-09; corroborated by Nakagawa (2024: 48–49)) ‹corpus-suggested›.

In some northeastern dialects, the accented vowel is slightly lengthened, an observation that underlies earlier romanization conventions such as ユーカラ for yukar (accent on yu, hence formerly written with a long vowel: Nakagawa 2024: 26, reported via Nakagawa (2024: 26)). Phonemic vowel length is not otherwise contrastive in Hokkaido Ainu; the vowel inventory is discussed in Chapter 17 (The Vowel Inventory and Vowel Realization).

23.5 Lexical exceptions

A set of words with open σ1 carry accent on σ1 despite the prediction of Rule 2 (open σ1 → σ2-accent). These must be stored in the lexicon; they cannot be derived from the placement rule. Nakagawa (2024: 51) lists the following as the principal exceptions in Saru/Chitose:

cápe 'cat', húci 'grandmother', húra 'smell', húre 'red', káni 'I (free pronoun)', kéra 'taste', míci 'father', mína 'laugh', nína 'gather firewood', nísap 'suddenly', núman 'yesterday', páse 'heavy', sápo 'elder sister', sések 'hot', síno 'really', sísam 'Japanese person', túki 'cup', túnas 'fast', úsey 'hot water', yúpo 'elder brother'.

Several exceptions pair minimally or near-minimally with regular-rule forms:

(Nakagawa (2024: 51); Shiraishi (2022: §5.1).)

The exceptions are not random: each corresponds to a Sakhalin Ainu form with a long first vowel (see the following section), and as a group they reflect the older vowel-length contrast that has been restructured as lexical accent in Hokkaido. A quantitative analysis of the exception set and its contrastive residue against the rule-conforming majority is given in Chapter 25 (Lexical/Contrastive Accent and Minimal Pairs).

23.6 The Sakhalin vowel-length correspondence

Hokkaido Ainu pitch accent stands in a systematic correspondence with Sakhalin Ainu vowel length: a Hokkaido word with accent on σ1 (including exceptions to Rule 2) has a long vowel in σ1 in Sakhalin; a Hokkaido word with accent on σ2 (regular Rule 2 output) has a short vowel in σ1 in Sakhalin Nakagawa (2024: 27, 51); Nakagawa & Fukazawa (2022: §3.1).

(SA forms from Nakagawa (2024: 27, 51).)

Hattori (1967, reported via Nakagawa (2024: 27) and Nakagawa & Fukazawa (2022: §3.1)) interprets the Sakhalin vowel-length opposition as diachronically prior, with Hokkaido pitch accent having developed from it; the diachronic relationship is argued in detail by Itabashi (2001) (not consulted directly). Satō dissents on the direction of the change ((Nakagawa & Fukazawa 2022: §3.1)) ‹contested›. The Hokkaido accentless varieties listed in the preceding section represent a third path: they have lost the length opposition without acquiring pitch contrast, and are not converging on the standard-accent system (Nakagawa & Fukazawa 2022 §3.1, reported via Nakagawa & Fukazawa (2022: §3.1)). Sakhalin vowel length and its relationship to the Hokkaido vowel inventory are discussed in Chapter 17 (The Vowel Inventory and Vowel Realization).