International Society for Knowledge Organization (ISKO)--North American symposium

The inaugural meeting of a North American subgroup of the International Society for Knowledge Organization (ISKO) took place on June 14-15 at the University of Toronto.  The theme of the symposium was “Knowledge organization research in North America:  What have we done, what are we doing, and where do we go from here?”

Brian Cantwell Smith, dean of the Faculty of Information Studies at the University of Toronto, noted that the current transformation from the print world to the digital world is on the same scale as the transformation from the aural world to the print world.  It was fitting that Richard Smiraglia, editor of ISKO’s quarterly journal, Knowledge Organization, announced that the journal would soon be published online; indeed there are plans to put up all issues of Knowledge Organization and its predecessor, International Classification, on JSTOR.

During the business meeting the assembly voted to form a North American chapter of ISKO.  A three-person steering committee (Kathryn LaBarre, Richard Smiraglia, Joe Tennis) will draft a set of by-laws for submission to the general assembly at next year’s ISKO conference and will begin planning for a 2009 conference of the North American chapter.

In addition to the business of voting to form a North American chapter, the symposium included the presentation of 13 papers and a panel discussion (James Turner, Clare Beghtol, Jens Erik Mai).   Many of the themes that were touched on have relevance for the Dewey Decimal Classification.

A recurring thread centered in the tension between traditional knowledge organization (KO) systems, such as Dewey, and the phenomenon of social tagging.   One manifestation of this thread is the question whether there is a single “right” way or multiple “right” ways to organize information resources.  Sometimes divergent classifications are compatible:  They may differ simply in the level of specificity they provide for a given domain, or they may recognize similar facets, but prefer different citation orders.  Mappings between such systems would be relatively easy to construct.  Sometimes a classification system has developed in a hodge-podge fashion and needs a thoroughgoing analysis and overhaul before its potential for compatibility with another KO system can be assessed.  But sometimes, classifications just are incompatible.  Social tagging usually doesn’t have an explicit classificatory basis, but some degree of classificatory structure may be discerned in the set of tags that a given person or a given community uses.   Although little of the requisite research needed to assess how social tagging is like or unlike traditional systems has been performed, we fear they will commonly be found to be incompatible.  If it is the case that end users classify a domain in ways that are fundamentally different from how traditional schemes classify the same domain, how should the traditional KO systems respond?  By redoubling our efforts to educate end users on how to use the tools the profession has developed for them?  By developing mappings between commonly used tags and classes/descriptors in the KO systems?  By adopting the perspective of the end user?  (Which end user[s]?)

Other insights:

1. KO systems generally need more semantics.  An important aspect of this effort is the recognition of a richer array of relationships.  (And our paradigmatic relationships need to go beyond hierarchy.)  Here the knowledge organization community shares some similar concerns with the natural language processing and Semantic Web communities, in contrast to the Web 2.0 community.
2. Facet analysis is one of the most important tools that the KO community has.  Information access and discovery might be improved by faceted search and navigation.  Are there ways of better leveraging the underlying faceted structure of legacy KO systems?
3. The KO community needs to focus its attention on using meta-analysis to build syntheses of multiple classification schemes.  In this quest for providing interoperability, one of the most difficult issues will arise from hierarchical inheritance in different systems.
4. The KO systems needed for organizing information resources may not be the same KO systems needed for organizing information.  Minimally, the latter will be at a finer level of granularity.
5. The KO community has not always succeeded in telling either itself or others what it is that we know.  Too often we have waited for others to ask us to organize information/ information resources.  With few exceptions, they didn’t ask, but often tried to develop such schemes on their own.  At best, they have rediscovered principles of knowledge organization.  The KO community needs to be proactive in developing rigorous classifications for emerging information discovery contexts.
6. The KO community may generally have more success developing smaller, local classifications than developing broader classifications.  But there is still a need for a framework to integrate such smaller classifications.
7. Conversation is often the context of learning and knowledge construction.  Can it also be the context for knowledge organization?  This point was illustrated with Wikipedia categories, the result of “conversations” among potentially many people.  It could also be illustrated with the DDC, with its circles of conversations, involving the editorial team, the Editorial Policy Committee, translation teams, user groups, professional organizations, etc.

Pictures from the conference (lots of them!) are available on flickr using the search term nasko_2007.