Derivation by Merge and Its Input Restriction
Hideharu Tanaka
July 2022

This paper explores the applicability of the structure-building operation Merge, addressing how its input search is independently restricted. The issue is discussed in light of an Anti-stacking Effect (ASE), which is a generalization subsuming a Case Adjacency Effect and states that no adverb phrase occurs above a head that shares matching features with an element at its edge. To derive the ASE in a principled fashion, it is proposed that derivation by Merge is constrained by a Matching Search Condition (MSC), under which if the head of one input X to Merge holds feature matching inside X, then the search for the other input Y is limited and cannot select any non-matching objects in the workspace. It is shown that the MSC deduces all instances of the ASE found in the domains of VP and CP, and even covers a crosslinguistic variation in the distribution of adverb phrases, once the varying height of verb raising across languages is considered. A notable implication of the entire theory is that Merge should only be constrained in terms of input restrictions and interface conditions, suggesting that no output restrictions internal to the computational system are required.
Format: [ pdf ]
Reference: lingbuzz/006723
(please use that when you cite this article)
Published in: Manuscript
keywords: merge, input search, workspaces, feature matching, case, adverbs, semantics, syntax
Downloaded:255 times


[ edit this article | back to article list ]