This article investigates the effect of 2 language-in-use factors on the introduction and maintenance of referents in instructive discourse. These factors, implemented as conditions in an instructive production task, were the assumed visual identity for the reader of the objects or referents to be referred to in the instructions (visually same vs. visually different) and the assumed goal of the reader (reading to do vs. reading to learn).The results show that both factors have a strong impact on the writers' referential behavior. Visually same referents are introduced and reintroduced fairly systematically by means of perceptually overspecified NPs. Visually different referents are introduced systematically by extra propositional identification speech acts and they are reintroduced more often by attenuated anaphoric expressions. Apart from that, writers show a number of referential strategies which fit in with the assumed reader's goal. Writers in the reading to learn condition use more overspecified expressions than writers in the reading to do condition.The results contribute to a more comprehensive understanding of the factors that determine the form of referential expressions in discourse. Referential expressions which are more informative than necessary for identification purposes do not only occur when the activation level of discourse referents is low, but also when the writer anticipates specific conditions in which the information will be used by the reader.