The Proceedings of AFLA 23 RE-LABELING “ERGATIVE”: EVIDENCE FROM FORMOSAN * Victoria Chen Shin Fukuda University of Hawai‘i University of Hawai‘i ychen6@hawaii.edu fukudash@hawaii.edu This paper examines the distributional restrictions on two basic case markers in morphologically conservative Philippine-type languages: (i) the morphological marking on the pivot, conventionally labeled “absolutive”/“nominative”, and (ii) the morphological marking on non-pivot external arguments, conventionally labeled “ergative”/“genitive”, and demonstrates that they are better analyzed as a marker of informational structure status (topic) and the reflex of structural nominative Case, respectively. With novel data from Puyuma, Amis, and Seediq, we present a nominative-accusative analysis for Philippine-type Formosan languages with an A’-agreement analysis for Philippine-type voice affixes, and argue for the presence of an A/A’-distinction in Philippine-type voice system. 1. Introduction * There is a consensus in the Austronesian comparative literature that a Philippine- type four-way voice system can be traced back to Proto-Austronesian, which is reconstructed as having the four-way argument-marking distinction presented in (1) (Blust 2015, Ross 2009, 2006, Reid 1979). 12 (1) A four-way case distinction reconstructable to Proto-Austronesian (i) Pivot: the morphological marking on the sole phrase in a clause eligible for A’-extraction (ii) X: the morphological marking on non-pivot external arguments (iii) Y: the morphological marking on non-pivot internal arguments (iv) Z: the morphological marking on locative phrases * This project is funded by Academia Sinica and the Linguistics department of the University of Hawai‘i at Mānoa. We are grateful to Atrung Kagi, Sunay Paelavang, Lisin Kalitang, Ofad Kacaw, and Dakis Pawan for sharing their languages, and to Edith Aldridge, Robert Blust, Henry Chang, Ting-chun Chen, Micheal Erlewine, Matt Pearson, William O’Grady, Yuko Otsuka, Stacy Teng, Shigeo Tonoike, and especially Dan Kaufman, as well as the audiences at NELS 46 and AFLA 23 for helpful comments on this paper. To remain theory neutral, we refer to the case markers reconstructed as ‘nominative’, ‘genitive’, 1 and ‘oblique’ in Blust (2015), Ross (2006), and Reid (1979) as pivot, X, and Y, throughout the paper. Aldridge (2016) makes a different proposal, claiming that the Philippine-type voice system 2 did not emerge after the split off of Rukai, a Formosan language that exhibits only an active- passive contrast synchonically. It is nevertheless uncontroversial that the four-way case distinction in (1) can be traced back to the ancestor of all Philippine-type Austronesian languages.