Chapter 19The Rhotic /r/ and Coda-r

19.1 The rhotic consonant

The consonant /r/ occurs both at the beginning and at the end of a Hokkaido Ainu syllable. A tap [ɾ] is its common realization, but speakers also produce trilled and other variants Nakagawa (2024: 28, 30); Satō (2008: 12). Its final realization deserves particular attention because a vowel-like release can make a closed syllable sound as though it ends in a vowel.

The phonological distinction between /r/ and /rV/ is supported by pronunciation, suffix selection, and behavior at boundaries. The release of final /r/ cannot be read as an additional vowel solely because it is audible Shiraishi (2022: §3.2).

19.2 Realizations of r

Nakagawa describes syllable-initial /r/ as usually a tap, with a trill or a voiced stop-like realization in some speakers. Shiraishi's review also reports lateral and [d]-like variants Nakagawa (2024: 28); Shiraishi (2022: §3.2). These are phonetic descriptions of one rhotic category in the cited varieties.

The precise realization therefore needs a speaker and context. Latin r identifies the phoneme without committing every occurrence to a single articulatory shape. A narrow transcription may add the detail required for a phonetic comparison.

19.3 Final r and its vowel-like release

The release of final /r/ often resembles the preceding vowel, giving the familiar term echo vowel. It can also be influenced by a following vowel or have a more central quality. Nakagawa discusses the pronunciations of pirka “good” and arki “come, plural,” while Shiraishi reviews instrumental and descriptive work on the variation Nakagawa (2024: 30–31); Shiraishi (2022: §3.2); Bugaeva (2012: 466).

A full vowel after /r/ can distinguish words. The pair etor “nasal mucus” and etoro “snore” is a standard example Nakagawa (2024: 30). In connected speech the difference can be difficult to hear, so grammatical behavior provides additional evidence.

Shiraishi recounts Chiri Yukie's observations, reported by Kindaichi, that /r/-final forms select consonant-following suffixes and participate in consonant sandhi. The nominalizer alternation illustrates the point:

The ending of mokor behaves as a consonant despite its audible release. This distinction was obscured by older spellings such as mokoro, as discussed in Chapter 14 (Historical Orthographies and the Batchelor Tradition).

Performance can make the release more prominent. Shiraishi describes lengthening in song and emphatic speech, as well as Satō's observation of a repeated final vowel used to fill a pause after /r/-final words Shiraishi (2022: §3.2). Such realizations need their prosodic context preserved in transcription.

19.4 Devoicing and frication in clusters

Shibatani describes devoicing of the rhotic after [k] and [p], devoicing with some frication after [t], and a fully fricated voiceless realization after [ʃ] Shibatani (1990: §2.2). These statements concern adjacent consonants. They should be distinguished from the vowel-like release of a rhotic following a vowel.

The phonetic variants leave /r/ in the consonant inventory. Their distribution can be investigated in recordings without treating each realization as a separate phoneme.

19.5 Changes at morpheme and word boundaries

Final /r/ participates in assimilation. Among the diagnostics reviewed by Shiraishi is the contrast between a final /r/ that changes before /n/ and a full /rV/ sequence that does not. The cited form a-kor-nispa “my husband” has a pronunciation with [nn] at the /r+n/ boundary, while poro nay “big river” retains the vowel before /n/ Shiraishi (2022: §3.2).

The environments, dialect differences, and degrees of regularity are treated in Chapter 28 (Assimilation and Cluster Simplification: Coda /r/, Nasals, and Clusters). This morphological evidence is also useful when an isolated recording leaves the presence of a full vowel uncertain.

19.6 Dialectal comparisons

Some northern Hokkaido records have alternatives with and without a vowel after /r/. Shiraishi discusses kikir ~ kikiri “insect,” citing Asai's observations Shiraishi (2022: §3.2). These forms show why the lexical shape must be established for the particular variety.

In Sakhalin, Tangiku describes a vowel following historical final /r/ and illustrates nukara “see” beside Hokkaido nukar Tangiku (2022: §2.4). A Sakhalin full vowel and a Hokkaido phonetic release have different structural status, even when their pronunciations seem similar. The comparative patterns are developed in Chapter 162 (Sakhalin and Northern Kuril Comparisons).