Improving accuracy is not the only reason for writing, and even if it were...

For research into language development in L2 writing to have any relevance, it has to be situated within a framework of decisions in writing pedagogy. Furthermore, a perspective on L2 language development cannot be limited only to accuracy levels. Even if this is the case, it is counter-intuitive that further input may be detrimental to language development, unless other factors are at work. Since Truscott's views cannot be ignored on the subject, some serious limitations in his more recent publications are identified, and his conclusions are later contrasted with four recent studies into error correction. However, before evaluating these latter studies, consideration is given to certain minimal features of what a communicative process writing approach and language correction within such an approach might suppose. Since it is relevant to the discussion, the significance of different aspects of input is discussed along with the issue of motivation, which Truscott has included in his thesis. Then, some research design and pedagogical issues are outlined since they are considered necessary to situate and analyse the four recent empirical research studies into error feedback on accuracy levels. With these studies in mind, a number of different writing task options are considered, which reflect possible pedagogical choices, apart from typical process writing. Finally, the conclusion is that the correction-effect-on-written-accuracy debate has actually narrowed the question of language development in discourse writing to levels of grammatical accuracy, and the recent research gives more of an appearance of including feedback on the third P of a PPP approach than anything else. (C) 2009 Elsevier Ltd. All rights reserved.