Chapter 11The Latin Phonemic Transcription

11.1 The phonemic principle

The Latin romanization used throughout this grammar is a phonemic transcription: one grapheme per phoneme, with no attempt to represent predictable allophonic detail. It does not attempt a phonetic (IPA-narrow) record of every surface variant, nor does it reflect the abstract underlying representations required for morphophonological rules. Where a predictable alternation is at issue — such as the realization of /s/ as [ɕ] before /i/, or the voicing of stops in intervocalic position — the transcription writes the phoneme and the allophonic detail is described in the relevant phonology chapter. The consonant inventory and its phonetic realization are treated in Chapter 16 (The Consonant Inventory and Its Phonetic Realization); the vowel system in Chapter 17 (The Vowel Inventory and Vowel Realization).

This transcription is the standard now used in Hokkaido Ainu descriptive linguistics and in the major reference works that underpin this grammar: Nakagawa (2024), Satō (2008), Bugaeva (2022), and Tamura (1996). The same system is employed in the present grammar's interlinear glosses and in all Ainu-language inline forms. The parallel katakana orthography is documented in Chapter 12 (Katakana Orthography and the Extended Small-Kana Codas); Sakhalin-tradition Cyrillic transcription, used only for contrast, is described in Chapter 13 (Cyrillic and Multi-Script Rendering).

11.2 Phoneme-to-grapheme correspondences

The inventory comprises five vowels and twelve consonants, listed in the table below alongside their IPA equivalents and the primary allophonic notes that bear on the transcription. Nakagawa (2022)

The system has no length distinction in vowels: Hokkaido Ainu does not contrast short and long vowels phonemically (a property that distinguishes it from the Sakhalin varieties, where vowel length is phonemic and corresponds diachronically to the Hokkaido word-final -h Itabashi (2001)). Hokkaido vowel sequences across morpheme boundaries are sequences of two distinct vowel phonemes and are treated in Chapter 22 (Glides, Vowel Hiatus, and the Diphthong Question).

11.3 The choice of c, and s before i

The grapheme c for the palato-alveolar affricate /t͡ʃ/ is the letter that has become standard in Hokkaido Ainu linguistics. Older transcriptions — including missionary-era works and earlier twentieth-century philological writing — employed ch, č, or occasionally ts for the same segment; these variants survive in reprints and in writers unfamiliar with the modern scholarly standard aynu-corpora Discord (2023–2026) (nukopoli, aynu-corpora Discord 2023). The unambiguity of c rests on a phonological fact: Hokkaido Ainu has no contrast between a palatal and a non-palatal affricate, so a single letter covers the full distribution of the phoneme without risk of collision with other segments Kirikae (1997).

Native-speaker writing, from Chiri Yukie Chiri Yukie (1923) to the Ainu kana orthographies documented by Kirikae Kirikae (1997) and Nakagawa Nakagawa (2006), does not write ca cu ce with contracted kana (チャ チュ チェ) but with the straight syllabary forms. Kirikae attributes this to the absence of any phonological contrast between a palatal and non-palatal series in Ainu: the affricate is uniformly palato-alveolar regardless of the following vowel, and the kana typography that distinguishes large from small character would imply a distinction that Ainu phonology does not make Kirikae (1997) ‹speculative›.

The fricative s is written uniformly throughout, even though its phonetic realization is [ɕ] (palato-alveolar) before /i/ and in certain coda environments. Earlier transcriptions sometimes employed sh or š for the coda or pre-/i/ variant; the modern standard writes s in all positions and treats [ɕ] as a predictable allophone Satō (2008); Nakagawa (2024). The phonetics of the alternation and its morphophonological consequences are developed in Chapter 16 (The Consonant Inventory and Its Phonetic Realization).

11.4 Glottal stop, vowel-initial words, and onset representation

Every syllable in Hokkaido Ainu has a consonantal onset. Words written with an initial vowel letter — ek 'come', arki 'come (plural)', opitta 'all' — have an underlying glottal stop [ʔ] as their onset, but the modern romanization does not write this stop. Tamura's earlier analytical papers represented it with an apostrophe ('a, 'i, etc.), and the Kotobai no kenkyū variant and some early descriptive materials similarly mark vowel-initial onsets; from roughly the 1980s onward, however, the apostrophe has been systematically omitted in published Hokkaido Ainu linguistics, the convention followed here aynu-corpora Discord (2023–2026) (nukopoli, aynu-corpora Discord 2023). The suppression of the apostrophe reflects a practical consensus rather than a settled phonemic analysis: whether the glottal stop is a separate phoneme or a default onset-filler is a genuine analytical question addressed in Chapter 20 (The Laryngeals: Glottal Stop, /h/, and the Final-h Question).

Word-final -h — which appears on many verbal and nominal stems in their citation form (e.g. nupurh 'mountain god' in some dialects) — is a distinct phenomenon: it is not a surface fricative but a laryngeal element that triggers morphophonological alternations and corresponds to Sakhalin vowel length. Its treatment is fully separated from the onset-glottal-stop question and is given in Chapter 20 (The Laryngeals: Glottal Stop, /h/, and the Final-h Question).

11.5 Pitch-accent diacritics: when and how accent is marked

Hokkaido Ainu has a lexically contrastive pitch-accent system. In citation forms and in paradigm cells where the accentual distinction is under discussion, an acute diacritic marks the first vowel of the accented mora: síne 'one', tú 'two'. When accent is not relevant to the point being made, forms are written without diacritics throughout. Running Ainu text — narrative examples, interlinear glosses, and ordinary in-text citations — is unaccented unless the prosodic pattern is the object of description; the pitch-accent chapters are the exception, not the rule Nakagawa (2024).

This policy mirrors the convention of Nakagawa (2024) and departs slightly from some pedagogical materials that mark accent more systematically. Most reference works in the literature — including Tamura (1996), Satō (2008), and Bugaeva (2022) — do not mark accent on forms cited in grammatical discussion, and the present grammar follows that tradition in its non-prosodic chapters. The accentual system, its placement rule, and the accentless-dialect split are developed in Chapter 23 (The Pitch-Accent Placement Rule and the Accented/Accentless Dialect Split).

11.6 Morpheme and person-marker boundary notation

Two punctuation marks segment the internal structure of Ainu words in this grammar and in the interlinear glosses documented in Chapter 10 (Interlinear Glossing, Abbreviations, and Citation Conventions):

• The equals sign = marks the boundary between a person marker and its stem: ku=kor '1sg.a=have', a=nu '4.a=hear'. This is an orthographic device of the Nakagawa lineage — expressly not the general-linguistic use of = for phrasal clitics Nakagawa (2024: 58–60): it keeps inflectional person marking visually distinct from derivation and makes the dictionary headword recoverable. The markers' actual affix-versus-clitic status is mixed and contested (see Chapter 69 (The Personal-Affix Template: Position Classes and Affix Ordering)).
• The hyphen - marks derivational morpheme boundaries and stem-internal junctures: yay-nu 'self-hear (listen)', e-nu 'appl-hear'. Derivational prefixes and suffixes form part of the derived stem, which the person markers attach outside of.

The distinction between = and - originates in Nakagawa's personal data-management notation of the early 1980s and entered print in the 1986 publication 『語りの中の生活誌』, from where it spread through Hokkaido Ainu teaching and descriptive materials Nakagawa (2024: 58–60); aynu-corpora Discord (2023–2026) (nukopoli, aynu-corpora Discord 2024). It has never been universal: Tamura's dictionary and papers print inflected forms solid or hyphenated, Refsing Refsing (1986) and Satō Satō (2008) use hyphens for all bound morphemes (ku-kor wa k-ek), and Ijäs Ijäs (2023) follows the hyphen practice; Nakagawa Nakagawa (2024) and the Handbook tradition Bugaeva (2022) use =, and it is adopted here. The morphological and syntactic status of the person markers is developed throughout Part V.

Word-division decisions — how compounds, incorporating forms, and postpositional groups are spaced — are a separate question from morpheme-boundary notation and are treated in Chapter 15 (Orthographic Standardization and the Word-Division Problem).

11.7 Capitalization, proper names, and punctuation

Ainu words are written in lower case throughout, including in the Ainu-language line of interlinear glosses. Sentence-initial capitalization does not apply to the Ainu line; only the free-translation line follows ordinary English sentence capitalization. Personal names and place names are capitalized when they appear in English running prose but retain lower case in glossed lines and in in-text Ainu citations: sisam mosir 'Japan (Wajin-land)', pirika menoko 'a beautiful woman'. Bugaeva (2022)

Punctuation within Ainu text follows the source when a text is being cited; the grammar's own prose interpolations and constructed examples carry no internal Ainu punctuation. The discrepancy between comma-period (，．) and Japanese-style pause-stop (、。) conventions in different Hokkaido Ainu publications is a document-production variable, not a linguistic distinction, and is not discussed further here.

11.8 Glide transcription and vowel hiatus

Sequences of two vowels from different morphemes — a=e '4.a=eat', u-epeker 'mutual.storytelling', i-omante 'send (with 4.o)' — are written as bare vowel sequences in this grammar, following Nakagawa's modern convention. An older style associated with Tamura's publications from the 1970s onward inserts a transitional glide at the hiatus site: uwepeker, iyomante aynu-corpora Discord (2023–2026) (nukopoli, aynu-corpora Discord 2024) ‹contested›. The choice has phonological consequences: whether the transitional [w] and [j] are phonemic or predictable epenthetic segments determines whether they should be written. The analysis adopted here treats them as predictable and omits them in the romanization; the underlying analysis is developed in Chapter 22 (Glides, Vowel Hiatus, and the Diphthong Question). Phonemic /w/ and /y/ in onset position are always written (e.g. wakka 'water', yuk 'deer').

11.9 Relation to the IPA and to other romanizations in the literature

The table below maps this grammar's graphemes to the IPA and to the conventions used in the four main reference works: Tamura Tamura (1996), Satō Satō (2008), Nakagawa Nakagawa (2024), and the Handbook of the Ainu Language Bugaeva (2022).

Across the four main sources, the segmental inventory is identical and the letter choices for the consonants and vowels are uniform ‹consensus›. The differences are notational conventions: the treatment of the glottal stop onset (suppressed everywhere in modern practice), the use of = versus - for the person markers, and the insertion or suppression of transitional glides at vowel hiatus. Tamura's 1996 dictionary uses only hyphens and inserts glides; Satō likewise hyphenates the person markers; Nakagawa and the Handbook use =; the sources are divided on glide insertion. These differences are confined to morpheme-boundary notation and to the analysis of hiatus; they do not affect the phonemic letter values.

The Shiraishi chapter in the Handbook provides a systematic phonetic description alongside phonemic notation Shiraishi (2022); where that chapter uses IPA symbols for phonetic detail this grammar cites the relevant IPA and maps it to its own grapheme. Older grammars and dictionaries — Batchelor Batchelor (1905), Chamberlain Chamberlain (1887), and the Kindaichi–Chiri tradition Kindaichi & Chiri (1936) — use substantially different letter choices that are not systematically enumerated here; readers encountering those sources should be aware that their ch and sh renderings are equivalent to this grammar's c and s, and that their morpheme boundary conventions differ throughout.

The relation between the Latin romanization and the katakana orthography — including the systematic letter-to-kana mapping, the extended small-kana coda letters, and round-trip conversion issues — is documented in Chapter 12 (Katakana Orthography and the Extended Small-Kana Codas).