The form in which some content, once selected by the speaker, is encoded in a sentence depends not only on the lexicon and the morpho-syntax of the particular language but also on the way in which the sentence is integrated into the information flow in on-going discourse. Regularities of this second type have for a long time played a marginal role in the history of linguistic research. If dealt with at all, they were considered to be a part of rhetoric.

To some extent, this has changed. In the 19th century Henri Weil (1844) was the first to state that languages tend to order constituents according to the principle 'from old to new' – thus relating a structural device, word order, to the distinction between old/maintained/given information and new information. Georg von der Gabelentz (1868) pointed out that a distinction has to be made between those constituents which say what a sentence is about ("psychological subject") and those which predicate something about the former ("psychological predicate").

Although their ideas were taken up by a few scholars, such as Hermann Paul, Karl Bühler and members of the Prague School, more systematic work on the "information structure" of a sentence only started a few decades ago. This research, ranging from phonology to formal semantics, has steadily increased over the years, and "information structure" has now become the subject of extensive research at many places.

It is surprising, therefore, that there is little systematic investigation of information structure in the learner varieties of children and adults. There is one major exception - the use of referential devices which differ in the degree of context dependence; thus, one and the same entity can be referred to by "an author –the author – Scott – he – 0", and the choice between these is partly determined by what has been said in earlier discourse. But information structure is evident in many other structural and lexical devices.

There are some related projects (outside the MPI).