Finding Subjects: Agreement, Case, & Articulated EPP The Problem:Minimalist analysis (e.g. Chomsky 2008, Richards 2012) assumes that T probes its c-command domain to acquire φ features and assign NOM Case to a DP which then merges into SPEC-T. Alternatively, EXPL fills SPEC-T (1a). However, there are significant disparities in φ agreement and Case marking possibilities between T and a c-commanded DP on the one hand (as evidenced in expletive sentences (ES)), and T and a SPEC-positioned DP on the other (1b-e). (1)a. There are books on the desk//Books are on the desk b. There is/?*are a boy and a girl at the door//A boy and a girl are/*is at the door c. There has to be hundreds of photos ... // Hundreds of photos have/*has to be ... d. There is only me in that picture // Only I am in that picture e. There’s books on the desk Thus, probing based on φ agreement and Case marking is not a viable mechanism for deriving Spec-T-positioned subjects. Related, and also ill-understood, is the phenomenon of ‘short movement’ toward but not to SPEC-T (2). (2)a. There arrived a train//A train arrived b. There is a train arriving (*a train) c. There was someone arrested (*someone) d. There was someone being (*someone) arrested (*someone) The Analysis: Φ agreement and Case marking are not central to locating subjects. EPP features alone locate and move arguments. Following the work of Richards & Biberauer (2005), Rezac (2006), Deal (2009) and others, EXPL is inserted low, into the empty SPEC position of non-inchoative unaccusative verbs (Deal 2009). Utilizing Chomsky’s (2001) observation that movement may have separate Agree and Merge aspects, EPP features are articulated into their Agree and Merge components, however, this articulation is lexically idiosyncratic—it is different for different verbalizing heads. v~ (‘default little v’), the verbalizing head for non-inchoative unaccusative verbs (e.g. arrive or be), bears the feature (3). (3) EPP for v~ : [ uThetaAGR, uDMRG ] (3) says that v~ must agree (≠ φ agree) a true (theta-marked) argument, but may merge into SPEC-v~ either that argument or EXPL. The latter possibility allows v~ of arrive to agree a train and merge there (2a). In (2b), the higher v~ of be must agree a theta-marked argument. If the lower v~ of arrive has merged there, locality only allows [uThetaAGR] of the higher v~ to access there, and the derivation fails. Therefore, in (2b), the lower v~ must agree and merge a train, so that the higher v~ finds the requisite argument to satisfy its [uThetaAGR], explaining requisite short movement. (2c-d) are similarly explained. T bears a Case-based EPP feature ‘[unom]AGR/MRG’ meaning that it assigns NOM to the DP in the SPEC of the immediately preceding phase (possibly there) and merges it into SPEC-T. In ES, the associate DP is assigned default Case (1d). There may bear gratuitous φ features 3/sg. If it does, φ agreement coincides with NOM Case marking (1c-e). If not, φ agreement must probe for the most local lexically concrete DP (1a-b). References Chomsky, Noam. 2001. Derivation by phase. In Ken Hale: A life in language, M. Kenstowicz (ed.), 1-52. Cambridge, MA: MIT Press. Chomsky, Noam. 2008. On Phases. In Foundational Issues in Linguistic Theory, R. Freidin, C. P. Otero and M.-L. Zubizarreta (eds.), 133–166. Cambridge, MA: MIT Press. Deal, Amy Rose. 2009. The origin and content of expletives: evidence from selection. Syntax 12: 285-323. Rezac, Milan. 2006 The interaction of TH/EX and locative inversion. Linguistic Inquiry 37: 685- 697. Richards, Marc. 2012. On feature inheritance, defective phases, and the movement-morphology connection. In Á. Gallego (ed.) Phases: Developing the Framework, 195-232. Berlin: Mouton de Gruyter,. Richards, Marc, & Theresa Biberauer. 2005. Explaining Expl. In The function of function words and functional categories, eds. M. den Dikken & C. Tortora, 115-153. Amsterdam: John Benjamins.