Wednesday, November 22, 2006

Returning to a Modest Proposal

I'm setting out to answer a question posed to me in an innocuous e-mail: "has anyone in the 'cognitive' camp taught you anything new about users? Are you influenced by anyone there?" This innocuous question is anything but. Instead of ripping off a pat answer to these pesky questions, I find that they have given me considerable opportunity to think about my current understanding of some of the core issues of library and information science. I'll address the questions in two different ways: in relation to my own research and in relation to my current employment as an industry analyst.

The mysteries of handbooks

My dissertation--a seemingly never-ending project--deals with a very simple topic: engineering handbooks. I began my career as a civil engineer and handbooks were a large part of both my education and my regular daily practice. Titles like "Concrete Design Handbook" and "Handbook of Steel Construction" still litter my book shelves. They are well-loved and dog eared, replete with various thumb tabs and margin notes. When I actually used them for engineering purposes they rarely sat on the shelf. Instead, they sat on my desktop within easy reach.

Engineers' use of handbooks is well known. Or at least, was well known. Some of the early survey-based studies of engineers noted that handbooks are very important sources of information to engineers. Then along camePinelli and his studies of NASA engineers. Handbooks were left off his questionnaires and became invisible as an important information source. Every researcher since has conducted their literature review due diligence and similarly neglected technical handbooks. From the perspective of needs and uses studies, they have become invisible.

I think my experience illustrates some of the problems with the "robo-techno" approach. The most pressing one is that if the object of interest doesn't appear in the questionnaire or the information system than it doesn't exist. This issue has, of course, been well bludgeoned by a number of other authors including Britain andBowker & Star. There is also another issue. Even if the object of interest is dealt with specifically in the studies, it may not be the most pressing issue. In general--and this is a huge sweeping statement--quantitative approaches are very good at determining what's typical, average, and mean. Phenomena beyond two standard deviations are simply fringe. Unfortunately, many of the most interesting aspects of information science lie in these weeds. It's no great trick to list the sources of information that are most commonly used by most people most of the time. Locating the unique, scarce, and uncommon is a far more difficult task. And yet locating a unique document that meets the needs of an individual in a specific circumstance is the most important consideration for librarianship (if you take Ranganathan's word for it).

This attention to the "other" is perhaps the vital contribution of the cognitive and social approaches within information science. Chatman focused exclusively on users that were outsiders. Dervin et al defined a path of research that turned away from systems and questionnaires and towards users. These researchers posited an approach that focused on the specific conditions of individual users: what they actually experienced, how they determined what information sources they required, and what motivated their pursuit for information. Upon reflection, I feel that Dervin's success--from a citation and keynote perspective--is primarily due to a few contributions to the tool chests of needs and users researchers: her critical incident/micro-moment time line inquiry method and her taxonomies of library "helps" and information-seeking "stops". These contributions have fleshed out many otherwise tired needs and uses studies.

I must now return to the question at hand: has anyone influenced my own research. In a word, no. I respect the work of both Chatman and Dervin but I find that their focus on users has a tinge of mania about it. The obsessive pursuit of detailing both the social conditions and mentalities of users (or non-users) freezes out one of the most important considerations for library and information science. And that consideration is documents, real things typically made out of paper and carrying some sort of textual or visual inscription. I have been influenced byFrohmann's recent call to "deflate information" and eschew some of the more ephemeral aspects of information science. From my perspective, the progeny of Salton's work focuses exclusively on the management of "information", a concept that I sometimes have a hard time wrapping my head around. I much prefer to stick to the study of actual documents, things that I can touch and smell if I really need to. Hence, handbooks.

My research on handbooks illustrates some of my concerns. I've already noted that engineering handbooks have been occluded in current LIS research. Unfortunately, common research techniques have been largely useless for capturing the richness of these enigmatic works. How can I possibly ascertain what handbooks represent to their users? Questionnaire-based research simply indicates that engineers use handbooks; I already knew that. Interviewing engineers using Dervin's approach leads to some very unsatisfying answers: "How did you find the information you needed to solve the problem?" "I used a handbook." "How do you use handbooks?" "[Frustrated] What do you mean... it's a handbook!" Even falling back on more traditional information science approaches has some extreme limitations. I conducted a study of reviews of handbooks written by engineers and that appeared in industry trade magazines. I wanted to capture how technical professionals actually described these works. The list of words that appear considerably more frequently than expected is far from informative: "handbook, piping, fluid, manufacturing, bookshelf, composites, CRC, anonymous, appendices." I realized that handbooks wouldn't give up their secrets easily.

I've since had to shift my focus. In order to study technical handbooks--a very common and banal creature--I've had to find a variety that is so truly odd and strange that I can see the individual threads of difference in the weft of ubiquity. I have traced the history of engineering handbooks back to the sixteenth century, a time when these works seemed very different. Surprisingly, some of the insights I've gleaned about early modern practices cast light on modern publishing juggernauts like John Wiley and Sons and McGraw-Hill. It seems strange that I've reached this destination by considering questions like "could Agostino Ramelli have worked for the Duke of Alva in the fortification of Artois during the 16th century Dutch Rebellions?" Obviously, neither Chatman nor Dervin would recommend such an approach but some of the authors they cite certainly do. I've had to move outside of typical LIS sources to consider the approaches of historians (particularly from the tradition of history of the book), sociology (including social critics like Foucault and de Certeau), and anthropologists (like Geertz). Am I still doing needs and uses? I think so but my committee may beg to differ. I'll let you know how it works out.

The Professional Perspective

Dervin and Chatman have had a far greater impact on my professional role than on my academic pursuits. I'm an industry analyst. I track various technologies and companies, write reports, talk to the press, and consult with our clients who may be considering implementing these technologies or purchasing from these companies. I am paid to be either a technology savant or an industry coroner, depending on the circumstances. At the end of the day, however, I don't actually do anything other than read, write, and talk on the phone.

Some of my colleagues think of themselves as consultants. I think a more apt description is that we're reference librarians: we consult sources and provide guidance. There is one difference between what we do and what a reference librarian does. Instead of relying on external resources for cognitive authority, we must create it. We are the experts and must engage in a variety of rhetorical practices to maintain our credibility. In our interactions with clients we can't cite sources (we are supposed to be the sources) and we give only the barest glimpse of our methodologies (since they're supposed to be rigorous and proprietary). We create a constant stream of reports that vacillate between tactical and strategic and contain all sorts of fancy visual tropes. The purpose of these reports is to provide guidance to our clients.

We constantly strive to improve our products and services. The notion of "improvement", however, is very interesting. It's difficult to apply common measures of quality when the corpus is in constant flux. We don't have the typical bibliographic luxuries such as classification systems, indexing, and a stable corpus. We don't even have a stable set of keywords since part of our business is forging brand new acronyms to describe emerging technologies. And we certainly won't reify the acronyms created by our competition. Regardless, management pursues a relentless quest to divine the needs of our clients from their search profiles. In many ways I feel that we're facing the tyranny of the MBA--reducing complex problems to a set of core metrics and five bullet points that are easily communicated at the next quarterly meeting. I know that these initiatives won't work. We are just repeating the practices criticised byBelkin and Dervin.

There is no way to reduce the complexity of the information problems that we are facing; they are far too rich. All we can do is gain insight about the behaviors of our users and anticipate their actions. It's difficult to bake insights into formal architectures and delivery mechanisms. We can, however, use the insights to structure the craft of what we do on an everyday basis: write and talk on the phone. When it comes to craft Dervin and Chatman are very useful. Instead of constructing mechanisms that enable our clients to locate the right document--whatever right means--we craft the documents themselves with Dervin's insights. We anticipate "stops" and build in "helps".

In short, I find the cognitive and social approaches to LIS research far too limiting for the research problem that I have elected to pursue. They are, however, very useful in my professional life.

And a final word

Some names are missing from what I've written, notably Cleverdon, Salton, and Swets. I have tremendous respect for the work of these researchers but I have to question whether their work is science. It is unquestionably brilliant but it may be more representative of engineering than science. Borrowing from Vincente, I want to suggest that engineering is primarily driven by methodology and consists of successively iterating through various design parameters to meet the requirements of a predetermined objective. Consider the Wright brothers tinkering with wings and bike parts until they created something that met their design goal of heavier-than-air flight. Science, on the other hand, is driven by epistemology and ontology i.e., establishing that a phenomenon is real and how that reality was verified. The cognitive and social approaches may merely be an attempt to import key concepts from other social sciences into information science. Instead of replacing one scientific paradigm with another, this movement could represent a shift from engineering to science. I don't mean to diminish the tremendous accomplishments of earlier systems-based work (I am, after all, still an engineer) but I do think that theLIS community needs to grapple with the entire realm of document use even if it involves seemingly bizarre side trips into the wilds of anthropology, nuclear physics, or historiography.