GIS as Social Practice:
or why standards are doomed to failure
Nicholas Chrisman
Geography Box 353550
University Of Washington
Seattle WA 98195 USA
Presented at UCGIS Annual Retreat, Bar Harbor Maine, June 1997
chrisman@u.washington.edu
http://weber.u.washington.edu/~chrisman
Outline of Presentation
- Evolution of my research
- (through presentations)
- 1987: Fundamental Principles of GIS
- evolution: Geog. of Geog. Info & Ethics paper
- What passes for 'debate'
- Social Practice of GIS
- Findings from studies of science & technology
- How context influences geographic databases and GIS technology
- Some incomplete examples (wetlands)
- Why content standards are doomed to failure.
Evolution of my research
1987: Fundamental Principles of GIS
p; GIS not built solely from geometric axioms
p; defined in institutional, social, historical context
p; develop needs by examining mandates
p; build cooperation by encouraging custodianship
1990: Geography of Geographic Information
p; cartography: neither art nor science (regulated utility?)
p; variations in production and consumption
p; importance of historical context
1992: Ethics paper (GIS/LIS)
p; auto-critique of cultural superorganic
p; role of disciplines as guilds
What passes for debate
Proponents of GIS
- Dobson: Automated Geography (neo-Hartshornian)
- Abler: macroscope
- Goodchild: GI Science (return to Bunge?)
- Openshaw: GIS as savior of geography (?)
p; technology as objective, unitary and universal
p; proof by demonstration (investigator as agent of technological change)
Critics (and other roles)
- Harley: Deconstructing the map (maps as power)
- Lake: GIS Wars (links to military power, rational planning)
- Curry: authorship, ethics ...
- Pickles: Ground Truth (GIS as virtual sign; Habermas ...)
- Harris & Weiner: (political ecology; GIS for South Africa)
p; technology as unitary and dubious
p; proof by demonstration (investigator as agent of social change)
Basic Question for GIS:
Where is "ground truth"?
Naive empiricism [heritage of Bacon, etc.]
testing accesses the "real world" (noncontingent truth).
Tacit dualism? [Plato/Kant lurking everywhere...]
belief in an underlying distribution approximated by repeated
measurement (truth hidden)
"Terrain nominale" (abstract universe)
[various origins at IGN: B. David, Salgé in Guptill and
Morrison ICA book; based on a more reflective philosophy of science?]
explicit recognition that a test only references reality through the lens
of a specification,
a set of rules; results contingent
The most empirical steps imply philosophic choices.
What do I mean by context?
Surroundings (milieu, environment)
provide systems of meaning
- social (organizations, groups, ...)
- cultural
- economic
- political
- disciplinary (organized ways of knowing)
plus the cumulative influence of each on the others.
Studies of Techology and Science
Impact of technology on society
Inexorable march of progress
Science inherent in Nature
Dialectics of ideas
Kuhn's paradigms: historical process of science
Scientific views contingent on frame of reference
Dialectics of power relations
Edinburgh school: science/technology play out political/economic/social
agenda of era
Social construction of facts
Detailed social process of stablizing facts
["Strong Program" of SSK: Latour, Woolgar, Pickering, etc.]
Stabilization of Facts
Latour and Woolgar (1987) Laboratory Life
present a process by which a scientific fact moves through 5 types:
from a speculation (hypothesis)
to a fact taken-for-granted.
The intermediaries involve degrees of qualification (modalities) that limit
the scope, specify who was saying this, etc.
Boundary Objects (Fujimura; Starr and Griesemer)
stabilize relationships between cooperators;
provide a stable object with multiple definitions
How context influences geographic databases and GIS technology
Databases
Disciplinary ways of knowing define specific measurement frameworks, models
of representation, tradeoffs between resolution in various components
GIS technology
not constructed in Redlands;
locally redefined by local power relationships
(some hegemonic, some cooperative)
Wetlands example:
vast differences in locating what is meant to be the same
Why content standards are doomed to failure.
Content standards p;
founding purpose of GIS: shared database
One producer will understand needs of many.
Standards can work
· to stabilize variation in practice,
· to consolidate agreement on procedures.
[in short to create black boxes that function in a larger system.]
BUT
standards will not protect against
· fundamental changes in the technology,
· lack of common ground (shared understanding).
How metadata changes the story
Metadata standards
change the roles; users must judge fitness.
Instead of packaging up "facts", the metadata provides a means
to participate in the construction process.
Metadata reopens the black boxes,
of course,
p; only when the metadata are reliable,
comprehensive & compreshensible;
p; only when the user takes responsibility.
Presented June 1997; web version Novemeber 1997