Semantic Web technologies provide a way to represent and to exploit data and can be particularly suitable for historical corpora. The Henri Poincaré correspondence corpus is composed of more than 2000 letters which constitute scientific, administrative and private exchanges. Several technologies have been used for this corpus: the rdf model, the rdfs knowledge representation language and the sparql query language. Recently, a navigation tool has been created to explore this correspondence corpus by exploiting similarities between resources. This tool can simplify corpus exploration and highlight unexpected relations between elements. It relies on the use of a flexible search mechanism based on the definition and the application of sparql query transformation rules. The system can be connected to any rdf database as long as underlying data is exposed through a public sparql endpoint.