Universal and Domain-specific Classifications from an Interdisciplinary Perspective

1
Universal and Domain-specific Classifications
from an Interdisciplinary Perspective

• Rick Szostak
• University of Alberta, Canada

2
Two Views

• A view (see the Leon Manifesto) which urges the
  development of a superior universal
  classification that would facilitate the access
  of especially interdisciplinary scholars to
  insights generated by all communities of
  scholars.
• A view that concepts are ambiguous and thus it is
  best to classify documents only within particular
  domains and in terms of concepts understood
  inductively as scholars within the community in
  question understand these.

3
My key argument

• These seemingly conflicting views are in fact
  complementary.
• The KO community can and should work toward both
  goals collaboratively.
• Interdisciplinary orientation guides a both/and
  rather than either/or perspective (though I was
  slow to appreciate the complementarity).

4
Pragmatic Approach

• It is useful to first ask what users need and
  then pursue this to the extent possible.
• Note that philosophers increasingly accept the
  need to combine philosophical and empirical
  analyses.

5
Interdisciplinary research is

• Critical to scholarship (Undiscovered public
  knowledge)
• Of increasing importance
• More dependent on information science than is
  specialized research
• Information science should guide scholars to
  relevant research they would not have looked for,
  by facilitating searches by phenomena and
  relationships. Note that most research
  investigates some sort of relationship.

6
Ambiguity of Language

• Often used as an argument for domain-specific
  analysis
• Is a question of degree
• Can be reduced by a universal classification
  grounded in a hierarchy of phenomena and
  relationships rather than disciplines
• Indeed, the way to reduce ambiguity is to reduce
  complex concepts to a set of universal phenomena
  and relationships.

7
One Big Question

• To what extent are complex concepts
• Only understandable in terms of other complex
  concepts
• Comprehensible in terms of a set of basic
  concepts that we might achieve consensus on
  across domains (phenomena and relationships)
• Domain analysis can be invaluable here.

8
Conversion

• Since scholars will wish to communicate across
  domains, some form of translation is essential.
• The best form of translation is a universal
  classification.
• Moreover, domain-specific classifications may
  find it useful to compare concept meanings with
  other communities.

9
Complementarities

• Induction and Deduction
• Ontology and Epistemology

10
The (Empirical) Way Forward

• Universal classification of basic phenomena and
  relationships use of linked notation for complex
  concepts.
• Domain-specific analysis of complex concepts in
  terms of basic concepts (not theories)
• Evaluate ontological and epistemological
  differences across domains
• Strive to reduce ambiguity