?Contextualization? is a central topic in Mode-2, a new form of knowledge production. It has recently argued that there are three criteria which mark how contextualization happens, and also three forms of contextualization that can be observed. This article investigates criteria and form in the Organizational Memory (KNEXT) of a Research Institute (IBV). Two main conclusions can be drawn: first, depending on the object of study, some criteria are more useful than others; in our case, located in a intraorganizational level, the second criterion -more concern with what we have called a strictly ?production analysis?- has more explanatory power. Finally, an evolutionary perspective on knowledge production, based on the appearing of Mode-2 objects in the KNEXT/IBV case, is outlined.