On linearization as part of narrow syntax. The case of OSV/OVS structures.

Abstract

In this paper I explore an approach to linearization where this is part of core or narrow syntax. Based upon Kayne´s (2011/2013) algorithm that processing follows the order of production, that is from left-to-right, and upon the timing condition that there be as little delay as possible between external Merge and eventual Spell-Out, I conclude that the constituent that appears linearized to the left-most extreme for any given sequence is the constituent that has actually completed the valuation of all corresponding features (Probe-Goal framework) before any other constituent. The evidence that I provide for such an approach relates to the accusative Case feature that O fails to value in (OSV) object-topicalization structures in English, and the accusative Case feature that O similarly fails to value in certain subordinate whstructures (OVS) in Spanish. I defend the theory that Merge applies bottom-up, as in standard minimalist accounts, though I point out that it is perhaps necessary to rethink the concept of (successive) cyclic movement.