Across languages, the morpheme expressing conjunction frequently has other uses as well. Several linguists have attempted to unify all uses of conjunction morphemes under one general algebraic scheme. We argue in favor of a more limited unification and propose a universal decomposition of conjunction structures: We propose that there exist both a “nominal” e-type and “verbal” or “clausal” t-type junctor. Our account is substantiated with evidence from synchronic typology and diachrony. Our analysis hinges on a generalisation, that e-conjunctors, but crucially not t-conjunctors, may have non-conjunctional quantificational meanings. Historically, we invoke the same principle to explain the change in the conjunction grammar of Indo-European which uniformly abandoned the e- and adopted the t-level conjunctions across the board.