Progress in geographic
information science requires a formal description of the ontology of the
geographic domain. Ontology has gained increased attention among researchers in
geographic information science in recent years, and can play an important role
in establishing robust theoretical foundations for geographic information
science in the future. Three broad sets
of foundational issues need to be resolved: (i) conceptual issues concerning what
would be required to establish an exhaustive ontology of the geospatial domain,
(ii) representational and logical issues relating to the choice of appropriate
methods for formalizing ontologies, and (iii) issues of implementation
regarding the ways in which ontology ought to influence the design of
information systems. It is widely recognized that the semantics of geospatial
information is critical for the development of interoperable geospatial data
and software. It is also widely recognized that GIS software and technology
should be able to interoperate with other software and databases such as those
involved in wireless applications, e-commerce, logistics, environmental health,
and health care delivery. Such interoperability requires a common or shared
ontology for the phenomena under consideration—any phenomena distributed
over part or all of the Earth’s surface. This means also that research in
the ontology of geospatial phenomena should be coordinated with efforts
designed to establish geospatial terminology standards.
An integrated approach
is necessary, because there is a strong interdependency between the methods
used to specify an ontology, and the conceptual richness, robustness and
tractability of the ontology itself. But such integration comes only at a
price. For while the potential advantages of ontology for the purposes of
information management are obvious, the task of providing a common reference
ontology (a single consistent and stable set of category labels) which would be
sufficiently rich to contain all the major taxonomical concepts used in all
scientific disciplines is, even when we restrict ourselves to the spatial
sciences, so enormous as to be unachievable without considerable compromises
along the way.
Ontology in the philosophical sense is
an enterprise that cross-cuts all branches of science and information systems.
This is because all sciences and information systems deal, in some way and at
some level of generality, with reality. There then comes into play what we
might call the ontologist’s credo: to create effective representations
it is an advantage if one knows something about the things and processes one is
trying to represent.
Ontological commitments, on the other hand, underlie all forms of cognition. Hence, ontology in
the information systems sense – ontology as the making precise of
conceptualizations and commitments – is needed to support research in
spatial information science insofar as the latter relates to how humans, both
experts and non-experts, use and understand geospatial software and theories in
spatial science. Ontology as the study of ontological commitments is thus close
to the research issues dealt with by geographic information science under
headings such as data modeling and representation.
Spatial reality existed even before
human beings entered the scene. The independence of the geospatial domain is
made especially clear when we reflect that there are several independent
branches of natural science – geology, geomorphology, pedology, climatology,
oceanography, ecology, forestry, and geography itself – that deal with
the same reality as that which is targeted by geographic information systems.
This independence implies, however, that the philosophical and the information
systems conceptions of ontology are much more closely allied in the geographic
domain than in some other domains.
The task of ontology building does not
focus on the design of specific algorithms and data structures that would allow
implementation and coding of geospatial information and processes. Rather it
aims to build robust, comprehensive and usable taxonomies. On this basis it
seeks appropriate representations for geospatial phenomena to match the
underlying reality, especially insofar as this reality is salient to human beings.
Submitted by:
David Mark and Barry Smith
Department of Geography
SUNY-Buffalo
Buffalo, NY 14261
dmark@geog.buffalo.edu;
phismith@buffalo.edu
