ABSTRACT
Compared to the concerns that have been expressed over protocol methodology in the last decade, comparatively little attention has been paid to issues of modeling. Even though the most complete and sophisticated model of writing, proposed more than 10 years ago, drew on the now suspect standard theory of protocol analysis discussed in the last chapter (Hayes and Flower, 1980a), little attention has yet been paid to developing a more appropriate alternative. Particularly if we accept the evidence presented in chapter 6 that protocols are a distinctive and routine product of participants' literacy practices, the question of how best to model these data becomes both legitimate and pressing.
