External set-Merge of heads and composite probing: the case of locative inversion in English and Chinese

The goal of this paper is three-fold. First, adopting Chomsky’s (2013, 2015) Merge-a model of syntactic derivation, I advance the hypothesis that composite probing (Coon & Bale 2014; van Urk 2015) is made available in English and Chinese to derive locative inversion via external set-Merge of C and T, forming {C, T} (see also Saito 2012 and Epstein et al. 2016). Second, I argue that English {C, T} hosts a non-selective A-probe, whereas the Chinese counterpart a relativized one. Importantly, I propose that this crosslinguistic contrast in the non-selective/relativized nature of a composite probe is deducible from the articulated structure in the left periphery and external set-Merge of heads. Third, what is crucial for the Merge-a model is the set of independent principles that regulate the generative capacity of free Merge by examining its varied outputs. I contend that the need of Case valuation on DPs, the activity of the φ-probe, and the Intervention Constraint all play crucial roles in English to exclude some structures built based on the introduction of {C, T} and its composite probe. Couched in the Merge-a framework, this paper suggests a new way to restrictively model the variation in A/A properties of movement across languages and across individual constructions language-internally. The proposed approach derives the relevant variation without stipulating construction-specific mechanisms or a variation in the nature of probes — the variation comes from the C-heads that merge with T prior to probing and how it interacts with independent syntactic principles.