There are three parts to this talk { related in rather tangential ways. First, I will give a recap of an argument developed in a couple of earlier talks { at IIiX in 2008 and at the SIGIR evaluation workshop in 2009. The gist of the argument is about thinking about IR as a science, and the consequences for both theory and experimentation in the eld. The second makes use of some ideas from general systems theory and the notion of open systems, and applies these ideas to information retrieval and the task context of search. The third discusses the status of queries in the current search world, and suggests two new ways to think about queries.