Liliana Sánchez' paper is a welcome contribution to the growing body of literature on Andean Spanish (cf. a recent survey in Muysken, 2004a), welcome both because a well-motivated and clearly described methodology is used and because it is embedded in an explicit theoretical framework. I do not have reservations about the overall conclusions of the study, but would like to draw attention to three issues: grammatical encoding versus pragmatic inference, the completeness of the analysis, and the issue of form learning versus meaning construction.