Memory texts (in this broad sense) can be regarded as interfaces between an individual and the past, but I prefer to treat them (in actor-network theory) as agents (actors) which interact with human agents (actors). Remembering is distributed between texts and other agents: neither operates autonomous, but they work together in a network. This networked or distributed remembering happens between one agent and one or more texts, as well as between several agents and several texts. Memory texts do not “speak for themselves” but only in communion with other agents. Let me give an example. A colleague asked me to recommend a book on collective memory. I knew the book I wanted to advise. I could “see” its color and its size in my mind. I even knew the name of the author, but could not recollect the title. Therefore I involved another agent: amazon.com. By entering the author’s name, the website yielded the title and a picture of the book, which I could then recommend. Who “remembered” the book? Neither I myself, nor amazon.com could remember in isolation: the two together were involved in a system of networked or distributed memory.