How universal is human conceptual structure? The way concepts are  organised in the human brain may reflect distinct features of  cultural, historical, and environmental background in addition to  properties universal to human cognition. Semantics, or meaning  expressed through language, provides direct access to the underlying  conceptual structure, but meaning is notoriously difficult to measure,  let alone parameterise. Using cross-linguistic dictionaries, we  provide here an empirical measure of semantic proximity between    concepts and analyse the structure of a network derived from it.  Across languages carefully selected from a phylogenetically and    geographically stratified sample of genera, translations of words  reveal cases where a particular language uses a single polysemous word    to express concepts represented by distinct words in another. We us  the frequency of polysemy linking two concepts as a measure of their    semantic proximity, and represent the pattern of such linkages by a  weighted network. This network is highly uneven and fragmented:    certain concepts are far more prone to polysemy than others, and there  emerge naturally interpretable clusters that are loosely connected to    each other. Furthermore, the networks of different language groups  exhibit consistent structures, largely independent of geography,  environment, and literacy. We therefore conclude the conceptual  structure connecting basic vocabulary studied is primarily due to    universal features of human cognition and language use.