Chapter 164Ainu Loanwords and Toponymy in Japanese

164.1 Overview

Ainu's classification as a language isolate, surveyed in Chapter 5 (The Genetic Position of Ainu: Isolate Status and Macro-Comparison Controversies), means that no genetic relationship accounts for resemblances between Ainu and Japanese vocabulary. Demonstrable shared items are contact loans, and the directionality of that contact is asymmetric by period and domain. The companion chapter Chapter 163 (Japanese Loanwords and Their Phonological Adaptation) documents the large stratum of Japanese vocabulary that entered Ainu after sustained post-medieval contact with the Wajin: domesticated animals, material culture, administrative titles, and, from the Meiji era onward, unmarked quotations of modern Japanese terms. The reverse current — Ainu material entering Japanese — is narrower in scope and less systematically studied. Its primary evidence is toponymic, and lexical loans proper are few. Vovin Vovin (2022: §1) observes that "it is silently assumed that there are no Ainu loanwords in Japanese" and acknowledges the absence of any comprehensive study; the same observation is echoed in Ijäs's typological overview of the language Ijäs (2023).

The geographical distribution of Ainu-origin place names testifies to a former Ainu presence across a wider area than the historically documented homeland. Shibatani (1990: §1) records that the toponymic distribution across Hokkaidō is extensive and that "many place names in the northern part of Honshū indicate that they might have once lived in that region as well"; Ainu place names also occur in southern Kamchatka. Chiri Mashiho devoted major scholarly effort to this toponymic substrate Chiri (1956), and that dictionary remains the standard reference for individual etymologies. Individual toponyms cited in this chapter as examples have been verified only at the level of their Ainu component elements through the Tamura and Nakagawa dictionaries, not against the primary place-name sources.

164.2 The toponymic substrate

A small set of Ainu geographic roots recur as diagnostic final elements in Japanese place names, constituting the backbone of the toponymic substrate. Because these elements are independently attested in the Ainu lexicon with stable meanings Tamura (1996: s.v. nay, pet, us, oma, to, sir, nupuri, kotan); Nakagawa (1995: s.v. nay, oma), the toponymic analysis rests on solid morphological footing at the elemental level even where the etymology of a specific place name requires verification against Chiri (1956). Japanese reflexes typically show rendaku (sequential voicing: pet → -betsu), phonological reduction, or orthographic regularization through kanji choice — adaptations governed by Japanese phonotactics, not by Ainu phonology Shibatani (1990: §1).

All etymologies in the table are illustrative; individual place names require verification against (Chiri 1956) and related primary sources. A specific Japanese name may involve phonological reduction, analogical replacement, or overlay by later Japanese morphology that obscures the Ainu source compound. The component Ainu forms are well attested in the living lexicon; it is the mapping from a given Japanese string to those components that calls for caution.

164.3 The nay ~ pet distinction

The contrast between nay ‘stream, tributary’ and pet ‘river’ reflects a live semantic distinction in the Ainu lexicon, not merely the presence of two suffix shapes. Tamura (1996, s.v. pet) characterizes it from the Saru area: pet designates the representative waterway of a locality — the Saru, Mukawa, and Monbetsu rivers all carry this label — while a branch stream or tributary is nay. The criterion is typicality, not size: even a small stream that defines its valley is pet, while a large tributary of a major river remains nay. Corpus frequencies reflect this asymmetry: pet occurs 1,344 times (rank 140) against nay's 559 (rank 312) in the Hokkaido text corpus, consistent with the broader referential range of the former.

The semantic distinction matters for toponymic analysis: a -nai place name designates a tributary or smaller mountain stream, while a -betsu name marks the primary waterway of its valley. A common folk-etymological error equates nay with 'pond'; nukopoli (aynu-corpora Discord, 2023-12-11, corroborating Tamura 1996) registers this as a mistaken conflation: 'pond' or 'lake' is to ‘lake, pond’, a semantically distinct root with its own toponymic reflex aynu-corpora Discord (2023–2026) ‹corpus-confirmed›.

164.4 The locative formatives -us(i) and -oma(i)

Two further formatives encode habitual or permanent location and are productive in place-name formation.

The verbal element us ‘be habitually at, grow at’ (Tamura 1996, s.v. us) means 'be habitually attached to, grow at'; the suffix -us(i) nominalizes this to 'place where X habitually grows or is found'. The compound ni-kap-us-i — ni 'tree' + kap 'bark' + -us-i 'habitually-at' — yields 'place where bark trees stand', commonly cited as the source of Niikappu 新冠 (illustrative only; Tamura (1996: s.v. -us(i))). The suffix also surfaces in Usu 有珠 and several other Hokkaido names.

The element oma ‘be located in/at’ 'be located in, be contained in' (Tamura 1996, s.v. oma; Nakagawa 1995, s.v. oma) forms the nominalized oma-i 'place where X is located', the source of the Japanese -omai / -oma endings. In Ainu syntax, oma also appears in postpositional constructions expressing containment; this overlap between its lexical geographic use and its grammatical locative function is discussed in Chapter 52 (Relational and Spatial (Locative) Nouns and Their Possessed Forms).

A further geographic compound illustrates the compounding pattern that underlies Hokkaido toponymy:

This compounding pattern — a landscape attribute or associated entity plus a geographic generic (nay, pet, to, etc.) — is the typical structure of an Ainu place name; the phonological adaptation that Japanese applies when borrowing such compounds is treated in Chapter 21 (Syllable Structure, Phonotactics, and Word-Edge Constraints).

164.5 The Tōhoku substrate and Vovin's Old Japanese evidence

The toponymic data extend beyond Hokkaidō into northern Honshū. Shibatani (1990: §1) observes that many place names in northern Honshū suggest historical Ainu presence there as well. Vovin Vovin (2022: §1–2) extends the argument into the philological record: he proposes that a number of items in Eastern Old Japanese — texts preserved in Man'yōshū books 14 and 20, compiled in the 8th century — are Ainu loans, reflecting a prehistoric Ainu substrate in the Kantō and Tōhoku regions ‹contested›.

Vovin also argues that modern Japanese iruka 'dolphin' is an Ainu loan, "isolated in Japonic; cf. Saru rika / Kuril rika 'whale meat'" Vovin (2022: §2) ‹contested›.

This programme forms a minority and contested position. Vovin (2022 Ch.6 §1) himself acknowledges the prevailing working assumption against Ainu loans in Japanese, and individual etymologies in the table are disputable on grounds of sound correspondence and comparative methodology; the full debate on Ainu's historical position and the validity of deep-contact proposals belongs to Chapter 166 (Macro-Comparison and Deep-Contact Controversies) ‹contested›.

At the community level, nukopoli (aynu-corpora Discord, 2026-01-22) proposes that the Iwate place name Otomonai (乙茂内) — both a surname and a stream name in an area dense with -nai endings — derives from Ainu o-tum-nay 'head-middle-stream'. The analysis is morphologically plausible within Ainu and consistent with the Tōhoku -nai substrate, but has not been verified against (Chiri 1956) or published place-name catalogues aynu-corpora Discord (2023–2026) ‹speculative›.

164.6 A case study: sikerpe and the Japanese siko-type

Ochiai (2026) provides a model of rigorous Ainu → Japanese lexical identification. The Ainu form sikerpe ‘Amur cork-tree fruit’ — fruit of the Amur cork tree (キハダ Phellodendron amurense) — is analysed as a reflex of Proto-Ainu *siw kor pe 'thing that holds bitterness': siw 'bitter' + kor 'hold, have' + pe 'thing'. The proto-form shifted to sikerpe through w-deletion and regressive vowel assimilation (o → e) Ochiai (2026: 196–199).

A cluster of Japanese dialectal names for the same tree — most conservatively sikorope in Akita, with reduced forms sikoro, sikonohe, siko no ki elsewhere — are loans from Ainu *si kor pe, borrowed before the Ainu vowel change to *si ker pe had completed Ochiai (2026: 199). Two complementary arguments support the claim. First, the Japanese siko- type is morphologically opaque in Japanese and has no Japanese-internal etymology, while the Ainu compound is transparent — the asymmetry that is diagnostic of borrowing direction. Second, the Japanese form preserves the earlier vowel /o/ in the second syllable, showing that the borrowing antedates the Ainu sound change; this gives a terminus ante quem for the Ainu historical phonology, linking contact evidence to reconstruction Ochiai (2026: 197–199). The reconstruction side of this analysis connects to Chapter 156 (Proto-Ainu Reconstruction: Segments and Accent Classes).

164.7 Lexical loans beyond toponymy

164.8 Methodological cautions

Place-name etymology faces several systematic pitfalls that bear on the analysis above.

Japanese phonological adaptation erases information present in the Ainu source. Coda consonants, the voiced/voiceless distinction absent from Ainu, and syllable-weight differences are all reorganized by Japanese phonotactics, as described in Chapter 21 (Syllable Structure, Phonotactics, and Word-Edge Constraints). Rendaku produces pet → -betsu; recovering the Ainu source requires knowledge of both the Ainu phonological inventory and the Japanese adaptation rules. Reading Ainu-derived place names through kanji orthography produces folk etymologies — 稚内 as 'young interior'? — that are wholly Japanese in structure and irrelevant to the Ainu source.

A second difficulty is the limited documentation of Ainu as spoken across many areas. As discussed in Chapter 4 (The Sociolinguistic Situation: Endangerment and Revitalization), recording of Ainu varieties was geographically uneven, and for areas with no documented variety — Hakodate, Erimo, Rumoi, Otaru — place-name reconstruction by extrapolation from better-recorded dialects is at best speculative. The constraint that a proposed source compound must be morphologically analysable within Ainu and that phonological correspondences must be regular remains the methodological floor for any specific toponym claim.

A third caution is convergence. Phonologically similar strings may arise from unrelated sources, and it is possible that some Ainu geographic terms entered Ainu at an earlier period from a continental substrate. Antitwilight (aynu-corpora Discord, 2024-11-15) has proposed Korean-origin candidates for several Ainu roots, including nay, casi, and others aynu-corpora Discord (2023–2026). This is unverified and has not been tested against the comparative literature; it is noted here as an acknowledged edge case, not a claim ‹speculative›.