Wordfish is a computer program written in the R statistical language to extract political positions from text documents. Word frequencies are used to place documents onto a single dimension. Wordfish is a scaling technique and does not need any anchoring documents to perform the analysis. Instead, it relies on a statistical model of word counts. The current implementation assumes a Poisson distribution of word frequencies. Positions are estimated using an expectation-maximization algorithm. Confidence intervals for estimated positions can be generated from a parametric bootstrap.The name Wordfish pays tribute to the French meaning of the word “poisson”.