Sequences of pronominal clitics in Mantauran Rukai: V-deletion and suppletion Doris Ching-jung Yen & Loren Billings National Chi Nan University This study investigates the fact that in Mantauran Rukai (an endangered Austronesian language spoken in southern Taiwan) a vowel in some pronominal-clitic combinations can be deleted but not in others. For each case, the attested combination results from interacting constraints on the morphophonology of pronominal combinations within an Optimality-theoretic model. In Mantauran, Zeitoun (1997b) suggests, NOM, OBL, and GEN pronouns are bound; a pronoun changes phonologically if it precedes an OBL pronoun. Hence, such clitic sequences form a morphological unit with fixed ordering. All OBL pronouns are /i/-initial in Mantauran. If a V-final pronoun precedes an OBL pronoun, then its final V is deleted (with one exception further below): =ʔo=inə ‘=GEN.2SG=OBL.VIS3SG’, =mita=iðə ‘=NOM.INCL1PL=OBL.INV3SG’, =ða=ilinə ‘=GEN.INV3SG=OBL.VIS3PL’, and =nomi=inamə ‘=NOM/GEN.2PL=OBL.EXCL1PL’. In these examples, underlining indicates deleted underlying segments. If the shape is /=...CV=i.../ (as exemplified above) or /=CVV=i.../ (i.e., =ɭao=inomə ‘=NOM.1SG=OBL.2PL’ or =nai=inə ‘=NOM/GEN.EXCL1PL=OBL.VIS3SG’), the pronoun-final V will be deleted: [=...C_=i...] and [=...CV_=i...], respectively; however, if the shape is /=V=i.../ (namely, with the preceding pronoun consisting of just a single V), then there is no deletion: =i=imitə ‘=GEN.VIS3SG=OBL.INCL1PL’. Our study investigates three complications: (i) the lexical GEN.VIS3SG allomorph /=i/, mentioned immediately above, where the V is not deleted; (ii) pronouns ending in VV, /=ɭao/ and /=nai/ above, delete only the latter V but not both; and (iii) suppletion in one (NOM.2SG) pronoun, with the C-final allomorph /=miʔ=/ only if it precedes an OBL pronoun, and V-final /=moʔo/ elsewhere. First, the GEN.VIS3SG variants /=ni/ and /=i/ are in complementary distribution: /=i/ merges with hosts ending in a velar stop (hereafter abbreviated as K) plus /a/; /=ni/ is used elsewhere. Two constraints require faithfulness to this lexical subcategorization in these special hosts, */…Ka=ni/, and the use of only the unmarked /=ni/ allomorph elsewhere, MKD/=i/. Two more constraints, in a markedness subhierarchy, prohibit pronoun-final VV and V if an OBL pronoun follows (where *VV]cl[cl.OBL » *V]cl[cl.OBL). We also propose a MAXX constraint prohibiting deletion of adjacent segments; it too is in a subhierarchy, dominating MAX (prohibiting deletion of any single segment). Finally, we use Kurisu’s R[EALIZE]M[ORPHEME] constraint, prohibiting the deletion of an entire morpheme. RM » *V]cl[cl.OBL (accounting for the nondeletion of a lone V in /=i/), and *V]cl[cl.OBL » MAX (accounting for the deletion of the final V in other combinations). Next, deletion of only the latter V in a VV-final pronoun is formalized by the ranking MAXX » *V]cl[cl.OBL » MAX. Finally, as for suppletion of NOM.2SG /=miʔ=/ and /=moʔo/, a MKD/=miʔ=/ constraint is proposed to select the right allomorph. The overall hierarchy is thus {RM, MAXX, *VV]cl[cl.OBL, */…Ka=ni/} » *V]cl[cl.OBL » MAX » MKD/=miʔ=/; also */…Ka=ni/ » MKD/=i/. The two subhierarchies above support these rankings as well. Beyond strictly formal issues, this study also makes a methodological point. Careful field research on a highly endangered language, resulting in quality documentation (by Li 1996/2004; Zeitoun 1997a, 1997b, 1997c, 2000, 2002, 2007; Zeitoun & Lin 2003—which our data come from), allows the wider linguistic community to see this theoretically interesting phenomenon. SAMPLE TABLEAU NOM [+me, –you, –PL]; MAXX *VV]cl[cl.OBL *V]cl[cl.OBL MAX OBL [–me, +you, +PL]. a. =ɭao=inomə * W * L ! b. =ɭa_=inomə * * c. =ɭ__=inomə * W L ** W The ranking of *V]cl[cl.OBL » MAX is established in a different tableau (not shown). In addition, deletion of the nonfinal V here (i.e., *=ɭ_o=inomə) would violate undominated CONTIGUITY (Billings 2002:66–68). REFERENCES Billings, Loren A. 2002. “Phrasal clitics.” Journal of Slavic Linguistics 10(1–2). 53–104. Kurisu, Kazutaka. 2001. The phonology of morpheme realization. Santa Cruz: University of California PhD dissertation. [Available from the Rutgers Optimality Archive: ROA-490-0102]. Li, Paul Jen-kuei. 1996/2004. “The pronominal systems in Rukai.” In his Selected papers on Formosan languages (Language and Linguistics Monograph Series C–3.1). Taipei: Institute of Linguistics, Academia Sinica. 415–441. Zeitoun, Elizabeth. 1997a. “Coding of Grammatical Relations in Mantauran (Rukai).” Bulletin of the Institute of History and Philology [Taipei: Academia Sinica] 68(1). 249–281. Zeitoun, Elizabeth. 1997b. The pronominal system of Mantauran (Rukai). Oceanic Linguistics 36(2). 312–346. [Zeitoun, Elizabeth] !"#. 1997c. “$%&' [Mantauran dialect].” In ()* [Paul Jen-kuei Li] (ed.), !"#$%&' [The Austronesian Languages of Kaohsiung County] (+,-./0123 456 [Kaohsiung Monograph Series 7]). +, [Kaohsiung]: +,-78 [Government of Kaohsiung County]. 159–297. Zeitoun, Elizabeth. 2000. “Dynamic vs. stative verbs in Mantauran (Rukai).” Oceanic Linguistics 39(2). 415–427. Zeitoun, Elizabeth. 2002. “Nominalization in Mantauran (Rukai)” Language and Linguistics [Taipei: Academia Sinica] 3(2). 241–282. Zeitoun, Elizabeth. 2007. A grammar of Mantauran (Rukai) (Language and Linguistics Monograph Series A–4.2). Taipei: Institute of Linguistics, Academia Sinica. Zeitoun, Elizabeth & Hui-chuan Lin. 2003. We should not forget the stories of the Mantauran: Memories of our past (Language and Linguistics Monograph Series A–4[.1]). Taipei: Institute of Linguistics, Academia Sinica. .