Probably the most interesting paper I've read in my one semester of library school so far has been Marcia Bates' 'The Invisible Substrate of Information Science'. Not necessarily because of the light it sheds on the field of information science, but because of the light it sheds on me, for being interested in information science.
One of Bates' basic assumptions is that information science is a meta-science; that is, as a field, it is concerned with the content of all fields. Academics and researchers within a particular field of science—particle physics, say—are only really concerned with the content of that field: major developments, the top researchers, the history of the field, and so forth. Librarians and other information scientists, however, must be able to organize, search for, and retrieve information from potentially every field and discipline; they must be able to use works of every style, content, and format, without necessarily knowing anything about the meaning of what they are organizing. (She identifies two other meta-fields—education and journalism—both of which also are concerned with the content of multiple other fields. And, like librarians, educators and journalists are also necessary for a healthy, functioning democracy. But hey, let's not get into politics yet, right?)
The comparison she makes—and a fascinating comparison it is—is to actors. In, say, a medical show, the best people to portray doctors are not, in fact, doctors themselves. This may sound obvious, but on reflection it seems absurd: wouldn't the person most able to act like a doctor is the person who does it all the time in reality? Similarly, the people most able to organize information are not those who create it—tell a bunch of biology professors to create a biological papers database or organizational system and they'll probably do a terrible job of it.
So—and not to sound too self-centered—what does this say about me? To be honest, my interests are diverse but my time is limited. I want to know about, say, high-energy physics, but I don't have the math skills (or the attention span) to study it in-depth. I'm interested in the political structure of ancient Japan, but studying several centuries of history is, you know, rather a large undertaking.
I want to know everything about everything—not necessarily because I think I can contribute something worthwhile to all these fields (I'm pretty sure I can't prove, say, string theory with just my pop-science knowledge), but just for the thrill of knowing. But if my expertise can help others—if, as a librarian, I can help the next Einstein find the books he needs—then I think I'll have gotten my money's worth from this degree.