As recurrently noted throughout the review of the developmental literature presented in the preceding chapters, some pervasive methodological problems plague the study of language and/or discourse development, resulting in divergent conclusions and claims concerning the timing, course, and determinants of acquisition. The first part of this chapter (Section 7.1) summarises these methodological problems, showing that studies differ along a considerable number of variables, making it difficult to compare and to generalise results. Taking this methodological discussion as a starting point, the remainder of this chapter (Section 7.2) then presents the rationale, methodology, and database of the cross-linguistic study to be presented in subsequent chapters, which was designed to address some of the unanswered questions previously raised.

Control of relevant variables

As noted previously (Chapters 4 to 6), the available developmental research on discourse development and/or on the acquisition of particular linguistic devices in several domains of child language presents some divergent results and conclusions concerning the factors that might determine acquisition. I argued that some of these divergences result from the different theoretical foci adopted by researchers across various disciplines and traditions. In addition, I briefly indicated some problems resulting from a pervasive methodological heterogeneity across studies, to which I now turn in more detail. I consider below different methods of data collection (Section 7.1.1), as well as a number of variables that require adequate control: discourse situations (Section 7.1.2), tasks and adult interventions (Section 7.1.3), materials and their mode of presentation (Section 7.1.4), and background knowledge conditions (Section 7.1.5).