This study explores compounds from the perspective of conceptual blending (conceptual integration), and argues that the meaning of compounds arises through the interaction of three levels: (i) input spaces established for the head and non-head components, (ii) a blended space involving compression and emergent structure, i.e. elements not imported from the input spaces, and (iii) the language system as a whole and the culture this system is part of. With regard to (iii) this study proposes the “Culture-to-Compound Hypothesis”, according to which compounding can be recruited to represent culturally “novel” content in languages where compounding enjoys a peripheral status in the language system. The examples discussed in the article come from Norwegian (a Germanic language where compounding is a central word-formation mechanism) and Russian (a Slavic language where compounding is more marginal in the language system).