Issues in the verbal morphophonemics of Ikema Ryukyuan

Phonology Forum Osaka University August 19, 2015 Issues in the verbal morphophonemics of Ikema Ryukyuan Yukinori Takubo Kyoto University [email protected]‐u.ac.jp Visit Kikigengo.jp for Ikema data Genetic relationship Japonic Japanese Nothern Ryukyuan Ryukyuan Southern Ryukyuan Macro Yaeyama Yonaguni From Pellard Thomas (2015) Miyako Yaeyama 2 1．Ikema Ryukyuan - Spoken in Ikema Island, Sarahama in Irabu Island and Nishihara/Nishibe in Miyako Island - Approx. 2000 speakers (fluent speaker are in the 50's or older) 北琉球語 Northern Ryukyuan 奄美諸方言 Amami Dialects 沖縄諸方言 Okinawa Dialects 宮古諸方言 Miyako Dialects 南琉球語 Southern Ryukyuan 八重山諸方言 Yaeyama Dialects 与那国諸方言 Yonaguni Dialects 宮古 Miyako 大神 Ogami 池間島 Ikema Island Irabu Island 佐良浜 Sarahama 西原 Nishihara Miyako Island 池間方言 I kema 伊良部 Irabu Miyakojima City Phoneme inventory • Vowels Short vowels /a,i,u,ɨ,(o),(e)/ Long vowels /a:,i:,u:,ɨ:,(o:),(e:)/  /o//e/ are used only for loanwords  /o:/, /e:/ are used for loanwords, responses (‘oo (Yes Hon.)’、’ee (Yes Plain’ ) and for a sentence final particle ‘doo (assertive)’ /ɨ/ appears only after /s,z,ts (f)/. There are minimal pairs with /i/ and with /u/. /siba/(worry : /sɨba/ lip ：/suba/(soba noodles /muusɨ /(burn):/muusi/(imperative of burn Consonants • Consonants /p,b,t,d,k,g,c,z,s,f,r,m,n,ɴ̥,ɴ,y,w,h/ • Geminates and consonant clusters /tt, kk, cc,ff,ss,zz,vv,rr/ Consonants with special distribution /ɴ/、/ɴ̥/ /N/: ɴm, ɴn, ɴ̥m, ɴ̥n, ɴɴ, ɴ̥N, Ns,Nz,Nt,Nd,Nk,Ng,Nf, (Nv) / /N̥/: ɴ̥mV, ɴ̥nV, ɴ̥N# /ɴɴ/(sweet potato) /ɴ̥N/(wear(shoes, pants),scoop(water) /ɴ̥na/ (rope) Transcription conventions • /ts/ c • /Nn/,/Nm/ nn, mm • /Nn/ n̥ n • /N/ n Syllable structure • C1C2 ｊ V(V)C3 • C1 is moraic • C3 is /N/ or one of the syllabic consonants : /s, z, c, f/ . The syllabic consonants are usually pronounced with [ɨ] e.g. [icɨ] (go). Basic vocabulary 1: Case markers • Genitive‐nominative ‐ga/‐nu • Accusative ‐u • Second accusative ‐a • Dative‐locative ‐n • Allative ‐nkai • Instrumental ‐hii • Limitative ‐taahii • Ablative ‐kara • Commitative ‐tu Focus particles • Topic =a • Focus =du • Focus+Nominative =nudu • Accusative topic =ugyaa • Additive =mai １． The basic structure of predicate (Hayashi(2013)) • [[verbstem+(CAUSative)+(PASSive)+(ASPect)]+(TeNSe)+(NEGation)]+ (MODal)+(TeNSe)+Sentence Final Particle Finite verb forms Basic forms (also for adnominal ) Class ic‐ɨ(go)/Class idi‐i(go out PAST icɨ‐tai/idi‐tai NEG ik‐an/idi‐n FUTure or Deontic icɨ=gamata/idi=gamata 2．Basic forms, NEG and SEQ Basic forms can also function as adnominal or indicative forms basic NEG • kak‐(write) kac‐ɨ kak‐an • kug‐(row kuz‐ɨ kug‐an • tat‐ (stand tac‐ɨ tac‐an SEQ (renyookei) kak‐i(‐i) kug‐i(‐i) tat‐i(‐i) (*taci) • ‘tac‐an’ cannot be accounted for. All the verbs the stem of which ends with –t‐ have this type of inflection. kit‐ win 、mat‐ wait 3.1 Verb classes and suffixes for causative/transitive vs. passive/intransitive Japanese verbs can be divided into vowel stem verbs and consonant stem verbs Japanese verbs Vowel stem mi‐ mi‐nai (see‐NEG) Consonant stem ik‐ ik‐anai (write‐NEG) The same applies to Ikema and other dialects in Miyako. • Vowel stem verbs have stem final /i/ idi‐ • Consonant stem verbs have stem final C kak‐ NEG CAUS PASS idi‐n idi‐ssasɨ idi‐rai kak‐an kak‐asɨ kak‐ai Problems of Ikema verbs that end in /u/ • In Japanese kau(buy) can be classified as a consonant stem verb that’s ends in /w/ kaw‐anai(buy‐NEG) kat‐ta(<kaw‐ta buy‐PAST) • In Ikema (and other Miyako dialects), /w/ dropped intervocalically. (kawa(skin)>kaa、 awa(bubble)>aa). • In Ikema, verbs that end in /u/, e.g. kau(buy)、kanau(work 、fau(eat), /w/ does not appear in the inflectional paradigm. NEG PAST SEQ kau(buy.basic) ka‐an kau‐tai Kai‐ • fau(eat.basic) fa‐an fau‐tai fai‐ Two possible accounts for the paradigm A) To posit /w/ (or some other consonant) in the underlying form. faw‐u>faφ‐u faw‐an>faφ‐an, faw‐i‐>faw‐i Problem: /w/ does not appear anywhere in the inflectional paradigm cf. sui(cut), tui get have /r/ in NEG and INF(initive): sur‐an(cut‐NEG),sur‐i‐ (cut‐INF) , tu‐ran(take‐NEG), tur‐i(take‐INF) B) To posit two inflectional classes Shimoji (2008) Inflectional class1 stem final segment is /i/ Inflectional class2 stem final segment is a consonant, /a/, or /u/ Problem： For some verbs their inflectional classes cannot be determined with their stem‐final vowels. 3.2 Inflection of verbs with /u/ ending stems mu‐u NEG umu‐un < umuw‐an CAUS umu‐usɨ < umuw‐asɨ PASS umu‐ui < umuw‐ai The paradigm can be accounted for if we assume umu‐ as a consonant stem verb and posit a progressive assimilation rule that changes /a/ to /u/, i.e. ua>uu. This assimilation rule can be observed in topic and accusative markers (zzu=a>zzuu (fish=TOP)) We will come back to this in section 5). 3.2 Inflection of verbs with a stem ending in /u/ Account B treats this verb as belonging to class 2 and assumes that it takes class 2 suffixes. assimilation NEG ‐an: umu‐an>umu‐un CAUS ‐asɨ: umu‐asɨ>umu‐usɨ PASS –ai: umu‐ai>umu‐ui Suruu become comlete 、ssuu pick up suru‐u VOL suru‐adi>suruudi NEG suru‐an>suruun CAUS suru‐asɨ>suruusɨ PASS suru‐ai>suruui ssu‐u VOL ssu‐adi>ssuudi NEG ssu‐an>ssuun CAUS ssu‐asɨ>ssuusɨ PASS ssu‐ai>ssuui • The assimilation rule is the same in Nagahama dialect of Irabu (Shimoji (2008) p288). Shimoji (ibid.) considers it to be the assimilation of a thematic vowel. /a/ in PASS and CAUS are not a themtic vowel. So the rule must be a pholological rule not mophologically conditioned. Arguments for analysis B • Reanalysis of class 1 verbs into class 2 Some class 1 verbs can be made into class 2 verbs by deleting stem final /i/ mui (get burned class1 /class2 • NEG mui‐n /mu‐un<mu‐an • CAUS mui‐ssasɨ /muusɨ<mu‐asɨ • PASS/POT mui‐rai /?muui<mu‐ai Ui(grow 、sui follow along 、hui get big Almost all the class 2 verbs can be made into class 1 verbs by changing final /u/ into /i/. Fau>fai (eat) PAST fai‐tai NEG fai‐n CAUS fai‐ssasɨ PASS/POT fai‐rai • The stem final segment of Class 1 verbs is to be underlyingly represented as /i/ . • The stem final segment of Class 2 verbs is to be underlyingly represented as consonants, /u/ or /a/ Apparent counterexamples of B umu‐ (think) can be made into class 1 by adding /u/ • NEG umuu‐n the same as /umu‐un(class 2) • CAUS umuu‐ssasɨ cf. umu‐usɨ • PASS/POT umuu‐rai cf. umu‐ui This case can be explained as an addition of /i/ followed by an assimilation of i to u: ui>uu. Apparent counterexamples of B fau(eat can be changed into a class 1 verb by changing /u/ to /a/. Fau>faa NEG CAUS PASS faa‐n (the same as fa‐an) faa‐ssasɨ faa‐rai Not a full‐fledged paradigm *faa‐tai, *faa‐suga This may be a case of ai>aa. ４．Verbs with a geminate stem cɨɨ(angle, wear) • Basic form cɨɨ • NEG cca‐n sɨɨ(know) sɨɨ fɨɨ ssa‐n ffa‐n • PAST cɨɨ‐tai sɨɨ‐tai fɨɨ‐tai • INF cc‐i ss‐i ff‐i fɨɨ((rain) fall) Hayashi(2013) Stem cɨ sɨ fɨ A) Posit Constraint (a): *ɨV Posit Rule (1) to avoid (a) ɨ‐i>ɨɨ Basic form ‐i cɨ‐i>cɨɨ sɨ‐i>sɨɨ fɨ‐i>fɨɨ B CC ɨ deletion rule: ɨ‐>φ /C＿V Geminates are created to preserve the number of morae Problem: No conditions are identified for the two outputs : CɨV>Cɨɨ, CɨV>CC Shimoji (2008‐p.69, p.282) Stem c‐ Basic form c‐ɨɨ Geminate Copy Rule c‐c‐an s‐ s‐ɨɨ s‐s‐an f‐ f‐ɨɨ f‐f‐an (/a/ is a thematic vowe) • Geminate copy insertion rule: if underlyingly moraic //C// and //(G)V// are adjacent in a word‐plus, then a geminate copy of //C// is inserted to produce a surface /CiCi(G)V/. (Shimoji p.69) Problem: GCI applies when a moraic consonant is followed by a vowel. /c,s,f/ are moraic consonants to which GCI applies. Basic forms are exempt from GCI. /ɨ/ in the basic forms must, therefore, be treated as not being a vowel. In Ikema, /fu/ and /ffu/ are both possible, making GCI nonapplicable. Table 6-10. Monomoraic stem and stem extension (Shimoji 2008 p.282） Example f‐ ‘bite’ s‐ ‘know’ c‐ ‘wear’ v‐ ‘sell’ ž‐ ‘scold’ r‐ ‘enter’ thematic ‐a f‐fa s‐sa c‐ca v‐va ž‐ža r‐ra thematic ‐i f‐fi s‐si c‐ci v‐vi ž‐ži r‐ri athematic f__ s__ c__ vv žž rr Takubo: geminate stems • Stem cc‐ • • • • cc‐an ss‐an ff‐an cc‐i ss‐i ff‐i cc‐asɨ‐ ss‐asɨ‐ ff‐asɨ‐ cc‐ai‐ ss‐ai‐ ff‐ai‐ NEG SEQ CAUS PASS ss‐ ff‐ (/a/ is not thematic) (/a/ is not thematic) • Basic form ‐ɨ *#CCɨ in this dialect Degeminate Rule applies CCɨ>Cɨ Cɨ>Cɨ: to meet the word minimality constraint ccɨ> cɨ>cɨɨ ssɨ> sɨɨ >sɨɨ ffɨ> fɨ >fɨɨ Historically, there is nothing that prevents CC stems. Geminate stems developped through assimilation of r to the preceding obstruents: sir>ss kir>cc fur>ff 5. Topic and Accusative forms in Ikema 5.1 The distribution of the topic /a/ 1. 2. 3. 4. Nouns ending in /a/: a ffa child +a ffaa Nouns ending in /u/: u zzu(fish +u zzuu Nouns ending in /i/ : replace i with jaa saki(酒 +a Nouns ending in diphthongs or long vowels: ja mai(rice +a maija 5. Nouns ending in Cɨ: Replace ɨ with CCa dusɨ(friend)+a > dus sa 6. Nouns ending in /n/: Add na iN sea, dog +a>inna sakjaa 5.2 The distribution of the accusative /u/ 1. 2. 3. 4. 5. Nouns ending in /a/, /u/: u ffa+u>ffau zzu+u>zzuu Nouns ending in /i/: replace /i/ with /yuu/ saki+u>sakjuu Nouns ending in diphthongs or long vowels: /yu/ mai+u>maiyu Nouns ending in Cɨ: replace /ɨ/ with Cu dusɨ+u>dussu Nouns ending in /n/: add na in+a>inna 5.3 Constraints and rules • The following rules/constraints can account for the distribution of the topic and accusative forms, if we assume the underlying form of topic form as /a/ and that of the accusative form as /u/ (cf. Celik(2013), Celik and Takubo (2014)). • 1)VVV constraint：*VVV VVja • 2)Diphthong constraint：*ia>ja、*iu>ju、*ua>uu • 3)Mora faithfulness constraint：*Ci+a>*Cja>Cjaa • 4) Morpheme boundary alignment constraint：*C‐a>C‐Ca 5.4 VVV constraints：*VVV VVjV • Nouns ending in a long vowel suu(winter melon)+a>suua suuja suu+u>suuu>suuju • Nouns ending in a diphthong mai(rice +a>mai a>mai ja inau(tornade)+a>inau a>inau ja mai+u>mai u>mai ju inau+u>inau u>ju 5.5 Diphthong constraint：ia>ja, mora faithfulness constraint • banti(we‐PL ‐a>bantia>bantja>bantjaa • banti‐u>bantiu>bantju>bantjuu 5.6 Morpheme boundary constraint • Assumption： Nouns ending in Cɨ are assumed to end in C, i.e. [ɨ] is assumed to be inserted phonetically cf. Shimoji (2008) . • in dog in a>in in u>in • umacɨ (fire) umac umac na in=a>i.na nu in=u>i.nu a>umac ca umac=a>uma.ca u>umac cu umac=u>uma.cu The morpheme boundary does not co‐inside with the syllable boundary unless C is inserted. • Cɨ can here be interpreted as CC underlyingly, given the argument in Section 4. It is possible with nouns ending in /z,c,s,f/ but maybe not for nouns ending in /n/. 5.7 ua>uu • Cu+a>Cuu • Progressive assimilation rule that (Celik(2013)). There is a phonotactic constraint that bans a falling diphthong in this language: *ia,*iu,*ua • NEG of umu‐u(think : umu‐un(<umu‐an think‐NEG 、CAUS umuusɨ (<umu‐asɨ), POT umuui(<umu‐ai) 6. *ia、*ua • In most (maybe all) dialects in Miyako, /ia/ or/ua/ is not allowed. Each dialect has different ways to get out of it. • Ikema、Irabu iu>juu ia>jaa, ua>uu • Ogami iu>ii ia>ee ua>aa E.g. mui(get burned) *mu‐as>maas (get burned+CAUS) • Kurima iu>juu ia>jaa, ua>oo E.g. Paku(box)+a >pakoo In Kurima the negative form of verbs undergoes different rules from that in accusative and topic forms. Kurima verbs kacɨ(write) kak‐an umu‐u(think) umu‐an>uma‐an • The leveling of /ui/, /ai/ ui>uu ai>ee is optional in Ikema mii‐rai‐ui>mii‐rai=juu see‐POT.PROG Summary • Verbal classes: we can posit two classes for the inflection/derivation of verbs: class 1 and class 2, which historically correspond to consonantal stem class and a vowel stem class, respectively. Synchronically there are reasons to think that the classes can be classified as –/i/ stem and non‐/i/ stem. • We can posit CC stem verbs to account for the CC~ Cɨ alternation. • There is a phonotactic rule to ban *ua, *ia, *iu in the dialects of Miyako. The particular ways out of it differ in each dialects. Residual problem • Phonetic account of CC~ Cɨ alternation Consonantal status of ɨ in other Miyako dialects. It can be interpreted either as a consonant or a vowel. e.g. pagɿ> [pagz] pagɿ+a>pagza If we take this position Hayashi (2013), Shimoji (2008) and Takubo comes much to the same thing. References • Hayashi, Yuka. 2009. Ryuukyugo Ikema hoogen no danwa siryoo (Discourse data of Ikema Miyakoan). Onishi, Masayuki, Kazuya Inagaki eds. Chikyuken gengo kijutsu ronshu 1, Gengo Kijutsu kenkyukai, 153‐199. • Hayashi, Yuka. 2010. Ikema. In Michinori Shimoji and Thomas Pellard (eds.) An Introduction to Ryukyuan Languages, Research Institute for Language and Cultures of Asia and Africa, Tokyo University of Foreign Studies, pp.167‐188. • Hayashi, Yuka. 2013. Minami Ryukyu Ikema hoogen no bunpou (Grammar of Ikema Miyakoan) Ph. D. thesis, Kyoto University. • Kenan, Celik, Yukinori Takubo. 2013. Saitekisei rion niyoru Ikema hoogen no keitai oninron (OT analysis of the morphophonemics of Ikema Ryukyuan, Chikushi Nihongo Kenkyukai, November 2nd, 2013. Kyushu University • Celik, Kenan, Yukinori Takubo 2014. Ikema topic and accusative markers, an OT analysis, FAJL poster sessions, NINJAL. • Shioji, Michinori. 2008. A Grammar of Irabu, A Southern Ryukyuan Language. Unpublished Ph. D. thesis. Australian National University. • Sugimura, Takao. 2003. Kurimatoo hoogen no kijutsuteki kenkyuu (Descriptive study of the Dialect of Kurima Island), Research report of Basic Research (C) (2). • Pellard, Thomas. 2009. Ōgami — Éléments de description d'un parler du Sud des Ryūkyū, Ph.D Thesis, EHESS. • Pellard, Thomas. 2015 The linguistic archeology of the Ryukyu Iskands. Published in Heinrich, Patrick, Miyara, Shinsho & Shimoji, Michinori 2015, Handbook of the Ryukyuan languages, Berlin: Mouton de Gruyter, pp. 13‐‐37. Corrected version.