Cross-linguistically, wh-movement (A-bar movement) is subject to constraints that have been challenging to describe precisely. One such constraint applies to movement from the complement clauses of verbs. Certain verbs, like shout, know, and regret, appear to block movement, as shown by unacceptable sentences like ??What did Mary shout/regret that John saw __?. In contrast, verbs like say or think, usually referred to as “bridge verbs,” do not do so, as shown by the acceptable What did Mary say/think that John saw __?.
This constraint is often argued to be related to pragmatics or discourse factors (e.g. Erteschik-Shir 1973, Ambridge & Goldberg 2008) and taken as evidence against the claim that constraints on movement are syntactic in nature (Ross 1967, Chomsky 1973, etc.). In this talk, I review the theoretical landscape, pointing out certain gaps in our understanding of bridge and non-bridge verbs. For instance, while there are hundreds of these verbs in a language like English, many accounts appear to be based on informal judgments for a much smaller number of verbs, and researchers do not always agree on what the appropriate generalizations might be.
Experimental syntax methods can help us close this descriptive gap and adjudicate between competing accounts. I describe results from an exploratory survey of 100 English verbs (cf. Liu et al. 2019). Statistical analysis suggests that existing proposals provide only limited coverage of the variation in acceptability for these verbs. In other words, movement is likely to be sensitive to fine-grained distinctions that are not fully understood, which in turn call for more systematic data collection as well as more robust theories of bridge and non-bridge verbs.