Experiments with discourse-level choices and readability

This paper reports on pilot experiments that are being used, together with corpus analysis, in the development of a Natural Language Generation (NLG) system, GIRL (Generator for Individual Reading Levels). GIRL generates reports for individuals after a literacy assessment. We tested GIRL's output on adult learner readers and good readers. Our aim was to find out if choices the system makes at the discourse-level have an impact on readability. Our preliminary results indicate that such choices do indeed appear to be important for learner readers. These will be investigated further in future larger-scale experiments. Ultimately we intend to use the results to develop a mechanism that makes discourse-level choices that are appropriate for individuals' reading skills