This dissertation explores the cross-linguistic nature of resumptive Ā-dependencies with a particular emphasis on resumption in spoken Arabic varieties. Two primary questions guiding this investigation are: (1) what properties do resumptive Ā-dependencies share with gapped Ā-dependencies and/or with base-generated binding dependencies; and (2) what properties do resumptive pronouns share with non-resumptive pronouns? The answers I provide to these questions are based on the most systematic investigation of connectivity and cyclicity effects under resumption to date, drawing both on novel data and on a broad survey of the previous literature. Regarding (1), I argue that two kinds of resumptive Ā-dependencies can be distinguished—those which behave like movement-derived dependencies, patterning with gaps (e.g. in Spanish), and those which behave like base-generated binding dependencies (e.g. in Arabic). The diagnostics which reliably distinguish the two types of resumptives and which march in lockstep cross-linguistically are morphophonological reflexes of movement, island-sensitivity, parasitic gap licensing, exactly stranding, and case-connectivity. A key corollary of my analysis is that natural language syntax must be able to differentiate between (External) Merge and Move (qua Internal Merge). To account for this distinction, I argue for a feature-driven approach to structure-building operations, as opposed to free (or untriggered) approaches. I adduce additional support in favor of the feature-driven approach coming from variation in the construction of long-distance dependencies, specifically with mixed chains. Regarding (2), I provide a battery of novel tests showing that resumptive pronouns behave morphologically, semantically, and syntactically like regular pronouns (Doron, 1982; Engdahl, 1982; McCloskey, 2002; Asudeh, 2004), regardless of whether or not the operator that binds them moves. Crucially, certain properties which have previously been taken tomotivate a movement analysis of resumption are shown to not march in lockstep with the aforementioned movement diagnostics: these are reconstruction effects and crossover effects. I contend that neither reconstruction nor crossover strictly correlates with movement, contra much previous literature, and I show that both are robustly attested with base-generated resumptive dependencies. Instead, independent structural properties of pronouns and general constraints on variable binding, respectively, predict the presence of both reconstruction and crossover under resumption. Additionally, I argue that the pronominal regularity of movement-derived resumptives (e.g. in Spanish) supports a stranding approach to such resumption, wherein resumptive pronouns double Ā-moved operators which bind them, over and against ‘spelled-out trace’ approaches which face significant hurdles in explaining all the relevant facts. This dissertation thus defends a pluralistic view of resumptive Ā-dependencies couched within a feature-driven approach to structure-building while nevertheless accounting for cross-linguistically stable properties of (resumptive) pronouns.