Knowledge base · Part V — Morphophonology and Sandhi · this topic as a network
29. Glide Epenthesis and Vowel-Hiatus Resolution
How does Hokkaido Ainu resolve vowel hiatus at morpheme boundaries?
The glide transition (わたり) and its non-phonemic status
How is the glide transition between vowels analyzed in Hokkaido Ainu?
In the Shizunai dialect of Southwestern Hokkaido, intervocalic -w- or -y- often appears epenthetically between an open syllable and a following vowel-initial syllable, both within words and compounds and between words in the same clause (Refsing 1986: 73). For example, tooka pe is generally pronounced [tookajpe] in that dialect (Refsing 1986: 101). In Chitose, when the prefix e- attaches to the verb ipe 'to dine', glide epenthesis occurs, yielding e-ype [ejpe] (Satō 2008: 210). The consonant y is likewise inserted when the reflexive prefix si- attaches to a vowel-initial form, as in si-y-osmak (Satō 2008: 249).
generated 4 claims ·
Glide epenthesis at the i- and u- prefix boundary
What conditions trigger glide insertion between a vowel-final prefix and a vowel-initial stem in Hokkaido Ainu?
Glides y and w are inserted to resolve vowel hiatus when the preceding vowel is /i/ or /u/, respectively (Shibatani 1990, Shiraishi 2022: 13.4). More specifically, when the prefix i- attaches to a vowel-initial stem, y is epenthesized between them (i- + V- → i-y-V-), and when the prefix u- attaches to a vowel-initial stem, w is epenthesized (u- + V- → u-w-V-) (Satō 2008: 13). The prefix i- obligatorily triggers insertion of y when it precedes the vowels e, a, o, or u (Satō 2008: 127), as illustrated by the adverb iyos 'after that', formed from os 'after, behind' with the indefinite prefix i- and an epenthetic y (Satō 2008: 292). Glide insertion is triggered only by a restricted subset of prefixes such as i-, u-, or si-, and does not occur after person prefixes like eci- or incorporated stems (reported in Shiraishi 2022: 13.4). In Southwestern Hokkaido (Chitose), this restriction is likewise reported: glide insertion occurs only after i-, si-, and u-, but not after eci= or incorporated noun stems (reported in Bugaeva 2012: 468–469). In the Southwestern dialect, glide insertion of /j/ occurs when the first vowel is /i/ and the second is /a, e, o, u/, with the glottal stop before the second vowel dropping (Bugaeva 2012: 468–469), while insertion of /w/ occurs when the first vowel is /u/ and the second is /a, e, o/, again with loss of the preceding glottal stop (Bugaeva 2012: 468–469). Bugaeva further analyses glide formation (/i/ > /j/ and /u/ > /w/) in the Southwestern dialect as occurring only when /i/ or /u/ is preceded by (C)V-type prefixes that assign accent to the following syllable, such as ku=, e=, i-, e-, ko-, or o-, but not after a= or eci= (Bugaeva 2012: 468–469).
generated 10 claims ·
The *CVVC constraint as a blocker of glide formation
How does the *CVVC syllable constraint affect glide formation at morpheme boundaries in Hokkaido Ainu?
Vowel weakening after personal and applicative prefixes
What process converts stem-initial high vowels to glides after certain prefixes in Hokkaido Ainu?
Stem-initial unaccented i- alternates to y- when preceded by a prefix, as in ko-ytak from ko- + itak, but does not become y- when the stem has inherent first-syllable accent, as in o'ika from o- + ika (Satō 2008: 13). The prefix i- alternates to y- before a vowel-initial verb only when preceded by a person prefix that places accent immediately after itself, such as ku- or e-, yielding forms like ku-yomare (Satō 2008: 14). Similarly, the prefix u- alternates to w- before a vowel-initial verb only under the same condition, yielding forms like ku-womare (Satō 2008: 14). Initial u- can also alternate to w- between a vowel and a consonant, as in e-wturentekehe from e- + uturentekehe, though with lower regularity than the i- → y- alternation (Satō 2008: 14). An unaccented stem-initial e is pronounced [j] after the person prefix a- (a- + e- → [aj]) (Satō 2008: 14), and an unaccented stem-initial o is pronounced [w] after a- (a- + o- → [aw]) (Satō 2008: 14).
generated 6 claims ·
Glottal insertion when glide cannot appear
When is a glottal stop inserted to break vowel hiatus in Hokkaido Ainu?
When glide insertion is not available — because V1 is a non-high vowel (e, a, o) or because glide insertion is blocked for morphosyntactic reasons — glottal stop insertion takes place instead (Shiraishi 2022: 13.4).
generated 1 claims ·
Vowel deletion as an alternative hiatus strategy
What is the role of vowel deletion in resolving vowel hiatus in Hokkaido Ainu?
Vowel deletion serves as an alternative strategy for resolving hiatus. When two identical adjacent vowels come together, they are reduced to one, as in kera + an → keran 'tasty' (Shibatani 1990). When two different vowels come together, the first vowel is normally elided, as in ine + an + kur → inankur 'which person' (Shibatani 1990). In forms such as hawean, haweokai, and haweash, the preceding vowel tends to drop when vowels occur in succession, yielding contracted forms hawan, hawokai, and hawash (Kindaichi & Chiri 1936).
generated 3 claims, 1 sentences without a claim ·
The phonemic status of the glide transition across dialects
How does the phonemic status of boundary glides differ across dialects of Hokkaido Ainu?
In the Shizunai dialect, a contrast was observed between suwat 'pothook' and suʔat 'pot bail', both analyzable as /su-at/, suggesting that the glide and the glottal stop are not freely interchangeable in that dialect (reported in Shiraishi 2022: 13.4). Also in the Shizunai dialect, glide insertion takes place in fossilized forms where the semantic transparency of the constituents has been lost, but not in semantically transparent combinations (reported in Shiraishi 2022: 13.4).
generated 2 claims ·