In English and many other languages, only some attitude verbs allow long-distance wh-movement from their declarative clausal complements. It is common to split the class of verbs into two sub-classes, “bridge” and “non-bridge,” such that wh-movement across the first class is acceptable (1) (taken to indicate well-formedness) and wh-movement across the second is not (2).
(1) Who does John say that Mary like __?
(2) *Who does John quip that Mary like __?
Complicating efforts to derive an empirically adequate characterization of bridge verbs is the fact that what counts as a bridge verb appears to vary across languages. In this presentation, I review two distinct hypotheses about bridge verbs that bear on the questions of how the bridge/non-bridge distinction is implemented syntactically and of how this distinction might be learned. In the first hypothesis, cross-linguistic variation in bridge verbs reflect the existence of an implicational hierarchy (Fodor 1992), which might have parallels in indexical shift and logophoricity (cf. Sundaresan 2018, Deal 2017). The second hypothesis does without an implicational hierarchy. Instead, bridge verbs within a language consist of verb classes that may or may not be (semantically) related to each other, and are learned in a way akin to how dative alternation is learned.