Corpora of English and Spanish were used to identify translation equivalents (TEs) for single instances of node words in the two languages with regard to the verbal environment of the node. Certain of these TEs were then used as nodes for new TEs in the original language, again according to the verbal context of the node. The extension of the resulting network of translation equivalence to give sufficient coverage of the languages involved was seen as the first stage towards automation of parts of the translation process.