In the fields of complex systems, theoretical biology, semiotic systems and cosmology, we find
convergence in laboratory techniques and theory related to mapping. To map something is to
understand it, both structurally and behaviorally. Mapping systems can be considered
complex automated semiotic systems animated by procedural specifications and constraints.
Physical and biological systems can be thought of holistically as a unified semiotic system. In
this paper I combine the subjects of two recent colloquia, one which reviews the work of
George Pattee, and the other on Mapping Knowledge Domains. The result: autopoietic
ontologies - self-organized maps.

Today we see a confluence of ambitious mapping projects in three historically distinct
domains: physics - mapping the inflationary universe (Guth); biology - the human genome
project; and epistemology - the semantic web. Currently, these mapping agendas have great
inertia in the domain of research and share many common methodological concerns due to the
fact they rely on software systems as means for organizing and representing the complex data
of their remote subjects. By remoteness, I mean that their subjects (territory) exist in the
macroscopic, microscopic or intangible domain of concepts (respectively), so that we must
rely upon what Michael Polanyi speaks of as tacit mechanisms for understanding them;
mapping of course is a primary strategy. These territories are, practically speaking, infinitely
complex and dynamic while their variables and multiple aspects can comprehensively only be
handled computationally.

New signaling channels emerge by constructing new ways of sending and detecting signals;
new observables appear by the invention of new measuring devices (Cariani). New
sensing/communicating/mapping/visualization methods have provided access to perspectives
we will never be able to physically access, such as observing the universe from an external
point of view, microscopic views, and apprehension of time frames that span the existence of
civilizations or nanoseconds. Maps are functional because they reduce in complexity and scale
to human perception that which they represent. But mapping has crossed over from mere
representation into modeling dynamic phenomenon. Maps are aware, have memory and are
responsive due to real-time sensing and algorithms which monitor their self-state.

Ontology, as a dynamic mapping experiment, may be automated by real-time world input and
self-organized according to principles of semiotic systems: autonomy, self-specification and
self-interpretation, while existing upon a materiality of signs and symbols. I will begin by
briefly describing the theory of complex semiotic systems, which is however covered more
thoroughly elsewhere throughout the literature, in order to enter a discussion on self-
organizing maps (Skupin).

A special issue of BioSystems regarding the work of George Pattee put forth papers
examining what is required in defining a minimal living system (Rocha, ed). Pattee feels that
physical and life systems do not operate by totally different rule systems, thermodynamics
versus developmental systems (description based reproduction), but are related. The articles
pay tribute to F. Varela's view of autopoietic systems while refining some of his concepts.
Varela thought that life is a systemic property, characterized by the autopoietic (self-creative)
way of organization, regardless of the components that come to realize it. Living systems are
guided by internal processes, recursive or reflexive. The artificial life community is indebted
to Maturana and Varela's views on emergence and self-generation.

George Pattee's work is distinguished by considering minimal living systems as based on the
principle of semantic closure, metabolism and individuality. Individuality is the principle of
spatial-topological unity, such as a cell membrane. Closed metabolic systems are constrained
by complex interactions of active agents that act on memory conditions (DNA), encoded
sequences which describe the initial conditions of the organism for reproduction. Pattee was
especially concerned with the relationship between biological/semiotic and physical systems,
which he referred to as the epistemic gap.

The epistemic gap is an inevitable situation which requires a distinction between observer and
observed; an observer that is capable of itself making distinctions between, at least, self and
not-self (environment). This situation is found at all levels of developmental systems. A key
position in scientific theory is given to the relationship between a system and the issues
involved in measuring or gaining knowledge of the system. Furthermore, the ability of
organisms to extend themselves (through the use of information or cooperative/symbiotic
strategies) to gain species advantage is what distinguishes a semiotic system from a
developmental system in general.

The term interactionism as used in the field of theoretical biology, has been refined through
iteration through cybernetics, second order cybernetics, autopoiesis, complex systems, and
alife (Oyama). In general, the concept of interactionism in respect to observer and observed
begins to define a system in semiotic terms. A semiotic system maintains that agents use
coded information - material symbols - to pass on heritable traits. The subject of phenotypic
transformation is of great interest and broad in scope. It is a subject that one might approach
on the hand from a biological study or on the other hand from a phenomenological (Varela) or
cultural/artistic study (Kristeva).

Closure or spatial-topological unity, a key aspect of a developmental system, gives rise in the
first place to the importance of spatial thinking in understanding phenomenon, and why
philosophers of metaphor have treated the subject of spatial metaphor so centrally (Lakoff).
The application of spatial metaphors to characterize information 'spaces' and networks
(cyberspace) is well established (Manovich). Since mapping strategies are implicit as a mean
of organizing and navigating spatial structures, mapping information or knowledge is a logical
step. As knowledge is a dynamic and emergent phenomenon, we can benefit from experience
in the field of biology which specializes in tracking emergent and evolutionary phenomenon
and combine this with cartographic techniques.

An important bridge from the principles of self-organizing systems to application of such
principles in the domain of knowledge modeling is found in recent research on semantic web
theory. An important meeting of the Arthur M. Sackler Colloquium entitled: Mapping
Knowledge Domains, was held at the Arnold and Mabel Beckman Center of the National
Academies of Sciences and Engineering in Irvine, CA, May 9-11, 2003. The forum addressed
topics of importance to semantic web research, such as "charting, mining, analyzing, sorting,
enabling navigation of, and displaying of knowledge" (Shiffrin).

We may speak of a semiosphere as the relationship of symbols of a scientific model to
correspondences in the world and an observer who makes this sense. Maps are generally taken
to have a one to one correspondence to what they represent, as opposed to mathematical
models which function not as mirrors but transformations of what they represent (Cariani).
Maps should reduce or exhibit scales of detail as their function is to make a mapped domain
graspable; a map the same size and detail as that which it maps is not of much use. Three parts
of semiotic functionality are measuring (sensing), computation (coordination) and effecting
(devices) (Cariani). Considering maps as semiotic systems then, the sensing aspect of the
system would be input methods (remote sensing, text input); coordination would be data
analysis and behavior algorithms; effecting is done through controlling devices (display,
rovers).

When we speak of 'navigating' information space, we are using spatial metaphor to understand
information structure. Three types of information topology can be described: relational,
topical, and structured. Each type of relation is a way of connecting (navigating) one point to
another. A relational structure is what Roy Ascott describes fondly in a recent article of
Leonardo, starting from the word 'Orai' while taking pleasure in the textual relationships that
one moves through. A topic map will group material by subject (Mane). Under the topic of
'coffee' one will find grouped aspects of agriculture, chemistry, coffee culture, and franchise
stores. An ontology map shows the structure of conceptual correspondences; ontological
components comprise the objects of a semantic network. An ontological relationship would be
Wave and its subclass SoundWave. SoundWave will inherit all the properties of Wave without
laborious re-specification, while adding a property which distinguishes it from its parent -
which is that this particular type of wave travels through a physical medium as a compression
front.

The field of Ontology has become the cartographic project underlying the semantic web
infrastructure. Knowledge domain specification is a cultural attractor. Practically speaking,
computerized ontologies arise out of knowledge management and metadata concerns. Typical
application areas are medicine and business automation. The Standard Upper Ontology project
(SUO) is an IEEE initiative to create a reference ontology for all others to append to and
inherit from - a metaontology.

The Digital Art Ontology (DAO) project is an online collaborative ontology authoring project.
However, since computational ontology is a new area of research, issues of practice have been
pushed to the forefront. Deliverables at this early stage are more in the area of software
systems, methods for collaboration and theory, and less about getting down to specification -
at least in the experience of the DAO community. Ontology efforts tend to encourage the
formation of communities on top of systems that can properly track semantic activity, as
humans are of course essential in guiding the knowledge base. It is assumed that an active
online ontology of digital arts would certainly be an important reference/aid to curriculum in
the digital arts, especially in the area of online education. In fact, as virtual universities evolve
freed from the constraints of modern/disciplinary campus architecture, it is hoped that the real
university itself will undergo ontologically inspired transformation, affording spaces for
semioticians, musicians and geographers to share classes.

The immediate goal of Criticalartware.net (CAW) is first to stimulate, then capture and map
discourse activity. CAW is concerned with creating a community as a way to sustain and
evolve its knowledge base. Underneath the hood of the interface is an XML map of relational
digital art terms, while the paths between nodes are themselves assigned a value indicating
fitness (well traveled). The generation of terms in the knowledge base, as it comes from an
involved, immanent perspective (members of CAW), avoids the problem of top-down
specification. Thus, most significantly (from the interactivist perspective) specification itself
emerges out of navigation (participation). One doesn't navigate a preexistent network of terms,
navigation itself creates/destroys territory or nodes as experience accumulates. With better
algorithms the knowledge base will begin to exhibit periods of stasis and critical
reorganization based on network tensions. Outside of such lived/design approaches,
knowledge terrains are highly artificial. Authoring ontologies in a more immersive design
situation will enhanced what Polanyi speaks of as an 'ontology of commitment.'

An ontology is an ecology of terminology which reflects the concepts in a specialized domain.
Strategies must be developed for cultivating locally well-structured concepts with agreed upon
methods for linkage. While the web generates the relational nodes that define the space of
terms, ontology is a further structuring activity, requiring ontological agents. The highly
interdisciplinary field of digital arts is the most challenging of domains for ontology, but is
also the field most equipped for the challenge of self-specification.

In his article, Cariani writes that "computation alone can't produce new empirical information"
(p.9). I believe that mapping can point to data gaps reflecting gaps in empirical phenomenon,
thus requiring new observables. For example, a missing term (concept) may be discovered
while translating ontologies form one language to another (a word for a species of snow).
Also, the navigation of an interactive map creates linkages and nodes that generate 'real' nodes
in the world in the form of emergent concepts and hence research topics.

Ontology is the age-old endeavor of naming and classifying and has always been involved in
mapping (surface) as a primary domain in itself. I refer to the issue of regarding representation
as somehow being of a secondary order - a second nature. Here lies the crux of the matter.
Ontology is a highly complementary cross-practice of creative design and field of analysis, the
interface where informing is occurring; perhaps it is the locus of interaction between genotype
and phenotype. Knowledge maps tacitly connect us to the phenomenologically distal
architectonics of the conceptual realm; they augment our ability to design and analyze,
combining the activity of the two cultures on an otherwise intangible canvas.

Ascott, Roy. 2004. Orai, or How the Text Got Pleated: A Genealogy of La Plissure du Texte: A Planetary Fairytale. In: Leonardo. MIT Press. Vol. 37. No. 3. p. 195.
Cariani, Peter. Towards an Evolutionary Semiotics: the Evolution of new Sign-Functions in Organisms and Devices. In Rocha, Ed.
CAW. 2004. Criticalartware. http://www.criticalartware.net/.
DAO: Digital Art Ontology. 2004. http://dao.cim3.net.
Gluth, Alan. 1998. Inflationary Universe: Quest for a New Theory of Cosmic Origins. Perseus Publishing.
Lakoff, G, Johnson, M. 1980. Metaphors We Live By. Chicago. University of Chicago Press.
Mane, Ketan K. and Katy Borner. Mapping Topics and Topic Bursts in PNAS. In Sackler Colloquium.
Kristeva, Julia. 1984. Revolution in Poetic Language. New York: Columbia University Press.
Manovich, Lev. 2001. Language of New Media. Cambridge, Mass: MIT Press. P.249.
Oyama, Susan. 2000. The Ontogeny of Information: Developmental Systems and Evolution. Duke University Press.
Petitot, Varela, Pachoud, Roy. 1999. Beyond the Gap. In: Naturalizing Phenomenology. Stanford: Standford University Press.
Polanyi, Michael. 1974. Personal Knowledge. Chicago: University of Chicago Press.
Rocha, Luis Mateus. Ed. BioSystems Vol. 60, No. 1-3. http://www.c3.lanl.gov/~rocha/pattee/
Rocha, Luis Mateus. 2001. The Physics and Evolution of Symbols and Codes: Reflections on the Work of Howard Pattee. In: Rocha, Ed.
Arthur M. Sackler Colloquium on Mapping Knowledge Domains. 2003. http://www.pnas.org/content/vol101/issuesuppl_1/index.shtml
Skupin, Andre. The World of Geography: Visualizing a Knowledge Domain with Cartographic Means. In Sackler Colloquium.
Shiffrin and Borner. Mapping Knowledge Domains: Introduction. In: Sackler Colloquium.
SUO: Standard Upper Ontology Working Group. http://suo.ieee.org/