As a graduate student at the University of Chicago, I occasionally sang with Andrew Abbott in choir — he was the bass in suspenders. It was only after moving halfway around the world that I began reading his work. I quickly became a fan. Abbott is one of the most thoughtful people writing today about what specialist knowledge is, and how we produce it. A historical sociologist with strong quantitative skills, he’s produced books on the history of academic disciplines and the dynamics of their formation and professionalization. But he’s also produced practical pieces about how students and professors develop ideas, and how to have new ones. There’s also an ‘applied’ dimension to his work — he produced the report on the University of Chicago’s library which made the bold move to double down on physical book purchases in what was supposed to be a digital future.

Abbott’s latest book, Digital Paper, continues this focus on the sociology of knowledge production by providing us with a “library methods” book: a ‘how to do fieldwork’ book, but for people who do library research. Andrew Abbott writing a book on how to do research? I was destined to like this book before I opened it up. But having read it now, and with a critical (if biased) eye, I can honestly say that every student, professor, and intellectual needs to read it. It’s a superb ‘how to’ guide about writing a long research paper or thesis. But it’s more than that. It’s an entire theory of how scholars pursue scholarship. It’s a memoir of Abbott’s own research. It’s a pessimistic and slightly misanthropic ode to a quiet world of well-ordered card catalogs destroyed by the garish vulgarity of online databases. It’s an epigrammatic summary of a career’s worth of knowledge. It is — yes, I really mean this — life-affirming. It improved my own ability to do research. Everyone needs to read it. You need to read it.

Digital Paper is organized into 12 chapters. Abbot takes nothing for granted — chapter 3 explains how to find books in the stacks using a call number, and how an academic citation is structured. Other chapters offer tours through more advanced topics, such as how to read a text (at what speed, and with what sort of focus) and how to write with style.  As a result, Digital Paper can help total novices learn their way around a library.

But for seasoned researchers, the book is also (as I said above) life-affirming. Easily my favorite chapter is the second, which provides an autobiographical account of the research that went into Abbott’s paper “Library Research Infrastructure for Humanistic and Social Scientific Scholarship in the Twentieth Century”. I had read this paper — which is superb — before reading Digital Paepr. Reading the story of how “Library Research” was produced was absolutely fascinating. In fact, I think if you just assigned the “Library Research Infrastructure” paper and chapter 2 of Digital Paper to students, you’d have a pretty good sense of Abott’s wider project.

But most importantly, the story Abbott tells in chapter 2 is thrilling. When he decides to check out every thesis written in the University of Chicago library school, you gasp at his audacity. When he unexpectedly finds biographies of great librarians while browsing through the Zs, your heart starts beating faster. The chapter is, in sum, a vindication for the obsessive research who lives in a world full of people who don’t care about bibliographic minutiae. Abbott reminds us that our insane desire for thorough and exhaustive knowledge of a topic is one of our most valuable traits — all that stands between the present and a future where the highest quality social scientific information available is Buzzfeed lists of celebrities who look like Otters. And for students who think we’re crazy when we say “Why don’t you read 400 things for your comps list and then decide on 30 to actually be tested on” We can now point to Abott’s book and say: This is how The Quality does it.

So the book has basic chapters for beginners as well as stories of expert research. But most of the chapters in Digital Paper, however, are well-pitched for the intermediate researcher, and move step by step through a research project: the preliminary phase of slowly getting a sense of what you are actually writing about, the mid phase process of developing bibliography on your project, the art of browsing the stacks, how to analyze material, and how to keep a file system you use to keep track of the project. Most of the chapters about the ‘mid phase’ of a project. The end phase receives relatively little attention because at that point, as Abbott point out, it is mostly a matter of assembling the paper out of bits of things you have already written.

On the one hand, Abbott makes a very abstract argument about how library research (and actually all research) works: It is ‘nonlinear’. By this he means that other manuals on how to do research are wrong: You don’t start with a lit review, then take notes, and then write up your paper. Rather, people are always already multitasking — as we spend time in the library or on the Internet we are silently engaging in all of these ‘stages’ of research simultaneously. As these processes cycle over and over, we feed them with material that sparks new ideas. As Abbott puts it, “serendipity is not an unusual, once-in-a-lifetime, even once-in-a-project thing. It is the one constant factor in library research.”

This abstract view of library research might be a little too schematic for the undergraduates who read Digital Paper in order to learn how to find books in the library. I think Abbott’s big-picture view of research will resonate most with professors and graduate students. But Digital Paper is not as dense (and has more signposting) than “The Traditional Future: A Computational Theory of Library Research”, where Abbott earlier made this argument. It is also much more concrete than “Methods of Discovery”, an earlier book of Abbott’s which zooms the camera out a bit too much. Digital Paper represents a plateau in Abbott’s writing on the subject of library research, and has hit the sweet spot between theory and accessibility. That’s one of the reasons that its so great.

One way in which Abbott helps bring this down to earth is his concrete instructions on how to find information. Many of us hate Google, but few of us hate it as much as Andrew Abbott, because few of us understand how search algorithms well enough to really be irked by them. ‘If you are looking for a needle,’ Abbot writes (I’m paraphrasing here), ‘why look in a haystack when you could go to a needle store’? By this he means that researchers should focus on high-quality sources, even if they don’t find everything (most things, Abbott believes, are probably not worth finding). Abbot prefers Web of Science searches (ranked by ‘most cited’) to Google. He’s also a fan of high-quality bibliographies, which can be out of date but still have a better signal to noise ratio than the Internetz.

I was skeptical about these recommendations, since my three biggest search engines are Google, Amazon, and Google. This is ok for me since I work in a small, young field (Melanesia), know most people in it, and I’m good at filtering noisy Google searches. So I tried a topic I’m interested in and don’t know very much about: Ernst Mach’s influence on Boasian anthropology.

Thanks to JSTOR, you can find and download Robert Lowie’s correspondence with Ernst Mach in a minute or two, and the Stanford Encyclopedia of Philosophy has an excellent entry on Mach… if you speak Philosophy. Mach’s writings are available on archive.org. So, yes: the Internet works and a tremendous amount of stuff is available… if you can make sense of it. But surely someone must have already written on this topic? There were scattered discussions of Mach (in Wittgenstein’s Vienna, Deacon’s biography of Elsie Clews Parsons, and so on), but nothing sustained.

But lo, using a World of Science search ranked by citation, I discovered a book review that led me to John Blackmor’s 2009 book Ernst Mach’s Influence Spreads which includes a chapter on his influence on anthropology. In the end, the chapter was not as strong as it could have been, but the process of finding it did help me understand how Abbott’s methods work for students who don’t have expert knowledge of the subject. So yes, I now do research differently because of Digital Paper.

Another major thing I learned from Digital Paper was the value of project-specific files. Abbott is (proudly) a bit a dinosaur and still prefers to work in paper, with a collection of folders for each specific project. He still does this even when he’s working digitally, and claims that after teaching his students the benefits of paper files, the scales fall from their eyes and they all begin taking notes with pen and paper. Initially, I was like: Well, that’s very sweet, and Abbott is entitled to his opinion.

But then I slowly realized that over the years, my library of PDFs had basically turned into a noisy intranet of its own and that my reading notes, although well organized, were ramifying out of control. So for my last article — on the influence of population biology on the thought of Jared Diamond — I tried Abbott’s approach of having a separate database for every file.

Again, I was happily surprised by how much better this worked than my usual methods. Creating smaller databases allowed me to find needles out on the Internet and organize them in my own needle shop, rather than just shoving them into the smaller haystacks that lived on my computer. I don’t think Abbott completely sold me on paper files, but he did sell me on project-specific files.

The problem with project-specific files is that each project makes sense, but your overall biography starts to lose coherence — you don’t anymore have a personal library or personal notes, its just one bloody project after another. But this is part of the sobering truth that Abbott conveys to us: Life is a coffee plantation. It’s your job, as a social scientist, to turn it into a cup of espresso. We are filters through which tremendous amounts of information pass, and in the end the final product of the research process is a paper which condenses and explains life — and a researcher who is a better filter than they were before. Scholars are not ‘learned’ in the sense that they know a lot, because we forget most of what we know. We hold on to it long enough to turn it into findings. Then we move on. As anyone who has ever written a piece knows, your readers often know more about your topic than you do since they read it more recently than you wrote it. From libraries we come, and to libraries we return.

There is a lot more to say about Abbott’s book — his advice on doing mini-analyses is very relevant for long term fieldwork, for instance — but in the end I’ll just conclude by saying that Digital Paper is truly something special. I can’t recommend it enough to students, professors, and researchers.