The starting questions proposed for our discussion of "information needs and information seeking" (as listed in the Borgman et al. background paper, June 1995) are: "To what extent are information needs and uses generalizeable across user and learner groups and to what extent do they need to be tailored? and "What is the relationship between information seeking and learning in digital libraries?" These questions are bracketed between guiding questions relating, on the one hand, to "social context and culture", and on the other, to "linking user(s)....to digital library design." For the former, the guiding question is: "To what extent can digital librar[ies]... be generalized across application domains and to what extent must they be tailored to each environment?" For the latter the question is: "What system design techniques are appropriate in applying user needs research to digital library design?" This set of questions is then followed with questions regarding the organization of information resources, the search capabilities of users, and interface design.
While the tone and thrust of the Borgman et al. background paper is light years ahead of what has become the normative approach to studying user information needs and seeking, I would like to propose that the questions are still grounded in the same meta-theoretic assumptions and that we need to jiggle these assumptions loose if we are to make significant progress in proposing as well as executing a research agenda that can usefully inform the design and implementation of digital libraries.
It is difficult to fully develop this issue in the space I have here. What I will do is attempt to circle around the issue showing the ramifications that the issue has for our project. I will then develop the ideas in more detail in February. To this end, I welcome queries, objections, and comments from readers which might inform my efforts to prepare myself.
Apologies in advance because I will seem to leave my assigned topic -- information needs and information seeking. I shall return to it, I promise. One of my main points is that we need to address the questions behind our questions regarding research agendas in such a way that we strive for a robust enveloping coherence. This may require that we strip away our attachments to certain ways of looking at the world -- to demography, for example, or user capabilities, or, even culture. Not that these ways of looking at the world are not potentially useful to information system design and implementation. But now we give them primacy without embedding them in a larger theoretically coherent whole. They become items on our far too lengthy and incoherently organized laundry lists.
The way our questions are set up we still look at: contexts, users, systems. Our attention is drawn to these as three discernible and separateable points. Only secondary attention is drawn to the connections between points. Hardly any attention at all to the threatening philosophic possibility that from the perspective of the user these are not three points, perhaps not even real moments of attention, perhaps only imposed edifices.
In a possible challenging re-framing, what becomes of interest is not context as such, or user as such, or system as such, but the behavings (internal and external) that happen to roam through these places so named (context, user, system) for academic/system convenience.
The difficulty we have is that while many fields have been calling, as we are now, for process-oriented and context-oriented studies, the results still end up primarily either as: incompatible mythifications convenient to academic or system design but often irrelevant to the micro-moments of human use; or as materially-anchored laundry lists without compatabilities across studies (as stated in the guiding questions of the Borgman et al. paper, across user groups, or across contexts). The issue becomes how can we theorize users, contexts, systems, and most important of all what happens in between (i.e. the movings, the verbings) coherently.
This requires that we re-examine the assumptions behind our quest for coherence. Of course, we seek predictability, we need it to make a system work. But the question we need to ask, the question behind the research agenda, is have we been looking for coherence in less than useful places. Traditionally, we have expected to find predictable patterns of relationships between user information needs and such characteristics of contexts as cultural differences and activity focus. Likewise, we have expected to find predictable patterns of relationships between user needs and such characteristics of users as their demography and personality. Mostly our research has failed us. The question that needs to be asked is: what if the coherence is not so much in these persons or places but in the movings and groovings in between? Might we genuinely theorize process and implement it in research practice as well as system design? Might this be a route for finding coherence and pattern where we now find diversity and disorder to be tamed.
To do this, however, requires that we dig deeper into the questions behind our research agendas to a very fundamental question about what information systems are about. When we focus repeatedly on such issues as "data/information validation" or "cognitive authority" as primary concepts of our effort, we may be inadvertently rebuilding philosophic traps. We collide here with issues of ontology and epistemology (and ultimately, of course, axiology, teleology, ideology). We still persist, I think, in conceptualizing the information system as a place where one gets something called information which at best is in some way isomorphic to reality. We still persist, I think, in trying to control and manage bias.
There is another possibility, however. One worth playing with. What if instead of trying to control and manage bias, we cherished it. What if we conceptualized every information "bit" as "biased" -- i.e. resulting from human observers, using human-created observational equipment and procedures at given moments in time-space, while moving along narratively-constructed journeys with historical pasts, presents, and potential futures? What if instead of hanging tightly to the dream of an epistemology whose certainty could approach a desired ontological certainty, we dreamed instead of a way in which we could capitalize on the creative ways in which humans persistently arrive at different answers to the same questions? What if instead of relegating these differences merely to epistemology, we reconceptualized them as bridges constructed in struggle and with reason across a quicksand called reality? What if instead of conceptualizing the different answers to the same questions as war, we re-conceptualized them as opportunity?
It is important to note that in the rambling above, what I have done is challenged us to stop making epistemology the culprit of our research scenarios [i.e. how can we find out enough about "them" to make our system work for "them"] and to start making ontology the opportunity. The ramifications would be enormous. We would not be designing systems as if every question ultimately had a "right" answer. We would put even the obvious into contest.
To do this, however, we would have to, in turn, reconceptualize some beloved shibboleths. Accuracy and validity, for example, would have to be reconceptualized as a few among the many evaluations humans bring to bear on their sense-makings of the world -- ways anchored in particular kinds of contexts and resulting in particular kinds of consequences seen as useful for particular kinds of journeys.
In such a framing, our questions turn from how can we tame contexts and users to how we can identify ways of looking at them that free them to be "real" [i.e. addressing the first of the two social problems identified for the workshop -- the need to identify "real" information needs"] while at the same time stripping away the unessential -- the "unreal" that has resulted merely from our own mythifications. This may require that we ask a very difficult set of questions: what is real about "users", about "contexts?" We may have to conclude, for example, that users are not real but what is real is the usINGS and the world viewINGS and the seekINGs and valuINGS that make up the usINGS. We may have to conclude that the idea of "user and learner groups" is not "real" and what is real is sometimes common, sometimes uncommon, and sometimes competing usINGS by real people on real journeys. [Emerson: Treat people well. Treat them as if they are real. Perhaps they are.]
At the same time we may have to incorporate two different additional dimensions as "real". One of these is an idea, currently relegated to impoverished conceptualizations of humans with such terms as "empowering users." What is impoverished about these conceptualizations is that because they are not tied to any materially anchored strategies for empowerment, they essentially relegate user to the babel of "others." The idea here that might be "real" is that the user has something of informative value to offer the information system. [For example, what might you teach your doctor about your body if: a) you had been taught to think about and observe your body; and b) your doctor had been taught that you had something of value to teach? What if such teachings by patients were routinely formally codified in the same way that we now codify the results of pharmaceutically-funded research studies?]
The second dimension which we may need to incorporate as "real" is the notion of iterativeness -- the idea that the system is designed to be always in process. The strength of the new technologies is that they allow this as a practical reality for the first time. What if, for example, a system automatically incorporated a way of storing for retrieval the last n challenges to an article? What if from among the last n challenges it was possible for weight to be given by subsequent user interest in the challenges? What if retrieval options linked to different strategies for evaluating (e.g. accuracy, harmony, tradition) and different strategies for answering (e.g. detailing, overviewing, illustrating) with the possibility always present for expanding on these? What if contest and consensus were both driving principles with links to each driving every moment of design?
A final ramification from now. In this reframing, we arrive at the idea that the core of the information system ought to be usings, not users. This implies a reconceptualization of all parties -- designers, practitioners, and users -- into carriers of usings.