This dissertation proposes a Construction Grammar account of how the meaning of complex constructions, specifically, analytic causatives, emerges from an aggregation of simpler constructions, which individually have fairly abstract semantics, but which in combination encode very specific event types. The constructions investigated are make s.o. V, have s.o. V, have s.o. Ving, get s.o. to V, force s.o. to V, set s.o. ( to) Ving, drive s.o. to V, move s.o. to V, lead s.o. to V, bring self to V, send s.o. Ving, give s.o. to understand, and leave s.o. Ving. Three causation event types are posited that are relevant to an analysis of analytic causative constructions: the MANIPULATE type, where an animate causer intentionally acts on a causee in a way that influences the causee such that he or she performs some activity; the TRIGGER type, where an event occurs which influences a causee such that, given the nature of the causee, the causee will inevitably undergo some process; and the PROMPT type, where an event occurs and a causee perceives this event and decides to react by performing some activity. Any given analytic causative encodes a more specific version of one or more of these event types. The two simple constructions that contribute most to a given analytic causative are causation verbs (make, have, get , etc.) and sentential complements (the bare infinitive, the to-infinitive, the present participle, etc.). Which event type(s) can be encoded by a given analytic causative does not depend on the causation verb or the sentential complement alone. Instead, particular combinations of the two yield meanings that are either compatible or incompatible with a given event type. Also discussed in the dissertation are the passivization of matrix or embedded clauses in analytic causatives, the relationship between causation verbs and their lexical sources, and the relationship between analytic causatives and transitive constructions with change-of-state or motion verbs.