A Constructivist Ontology for Structures and Functions
Aldo Gangemi and Giovanni Pezzulo
ISTC-CNR viale Marx, 15 – 00137 Roma
gangemi@ip.rm.cnr.it, pezzulo@ip.rm.cnr.it
We will describe an ontological perspective about Descriptions and Situations, and we will give
some cues to couple them with a theory of functions, mainly in an Activity Theory perspective. This
method is being used within a general computational framework for modeling Artificial Cognitive
Agents.
Descriptions and Situations in Formal Ontology
Foundational ontologies are being used in artificial intelligence and computer science as axiomatic
theories that contain a specification of domain-independent concepts and relations, based on formal
principles derived from linguistics, philosophy, and mathematics. Examples of formal principles are
spatio-temporal localization, topological closure, heterogeneity of parts, etc. The traditional formal
ontology field in philosophy is therefore partly reinterpreted according to new needs and
applications.
Non-physical objects, such as social institutions, organizations, plans, regulations, narratives,
mental contents, schedules, parameters, diagnoses, are a challenging issue. Many have negated an
ontological primitiveness to them, because they their intended meaning only results from a
statement. This position is documented in important theoretical frameworks (BDI, theory of trust,
situation calculus, formal context), that only consider states of affairs, facts, beliefs, viewpoints,
contexts, whose representation is set at the level of theories or models, not at the level of classes or
relations.
Here we support the position by which non-physical entities can be represented both as
theories/models as well as concepts with explicit reification rules. The Descriptions and Situations
ontology (D&S) is an attempt to define a theory that supports a first-order manipulation of reified
theories and models, independently from the particular foundational ontology it is plugged in.
In general, D&S commits only to a basic ontological distinction between flux, or an unstructured
world or context, and logos, interpreted here as intentionality, or description. D&S is neutral with
respect to realism issues, such as whether we conceive a structure, (called situation in D&S)
because it is already in the flux, or because it is created by our intentionality. Hence, a flux can have
as many inherent structure (parts, boundaries, qualities, etc.) as one might want to believe in or
might claim to have discovered, but without a description, no situation would emerge.
Due to its neutrality with respect to realism, D&S can generalize the flux/logos distinction, in order
to obtain an epistemological layering. This consists in assuming that any logical structure is built
upon a flux-like structure that it describes according to a more abstract, logos-like theory (either
formal or capable of being at least partly formalised). Epistemological layering reflects the figureground shifting cognitive process. Most assumption-making in any domain of interest seems to
apply epistemological layering (otherly named as tacit knowledge, context, substrate, etc.).
In order to be used for modelling cognitive agents according to the assumptions in the
Pandemonium metaphor, the basic ingredients (components) of descriptions are assumed to be
schema-like relations, in the vein of cognitive semantics, which suggests a “grounding” of its
schemata in human-environment interaction. In the following, the phrase schematic structure will
be used as “higher generality” descriptions.
Descriptions and Situations in Cognitive Modeling
Descriptions and Situations can be used to model the representational apparatus of a Cognitive
Agent. In a constructivist perspective, representations have to couple symbolic structures and
behavioral functions.
We assume the Pandemonium metaphor (as in Jackson, 1987) to model a Cognitive Agent that has
many isolated and concurrent processes (Demons), each carrying only a single operation and
applicable only to one kind of structure.
If we analyze the conceptualization activity by starting from perception, a stimulus is an open
problem for a cognitive system: many Demons try to interact with it, singularly or in Coalitions (for
more complex operations). A Demon that interacts with a stimulus cuts it from the noise; this
constructive operation results in applying a description and building a situation. For example a
Demon that carries a pattern matching operation only matches a certain pattern and it is constrained
to describe data in that way.
Representing is not mirroring the environment, but the constructive operation of fitting stimuli into
a schematic structure: the same constructive operation of perceiving a stimulus fits it into a
description and constrains how the system can deal with it.
Hence, situations pair intelligibility with action possibility. Problem-solving capabilities are
embodied in situations as their functional-effective counterpart; e.g. the constructive operation “putin-line” builds a situation that is associated to some “plots-to-put-in-line” and some “plots-to-dealwith-objects-in-line”: there is a functional link between situations and objects that can take a role
into, as well as there are constraints for the (structural) functions the system can apply to them.
Therefore, constructive operations contribute to build a cognitive map of the environment, fully
grounded in the internal, functional structure of the system.
The conceptualization process is both constrained by environmental stimuli and schematic
structures; moreover, it receives both contextual and motivational pressures, because descriptions
are underdetermined with respect to stimuli.
From an epistemological point of view, descriptions are theories that compete for explanation of
phenomena. Following Activity Theory, representational activity is “organizing for use”, but also
constraining it: an active structure, once instantiated, constraints the way stimuli can be
successively perceived and dealt with.
A related claim is that an organized situation is also a cognitive artifact affordable for anticipatory
and proactive mechanisms.
This process is not only bottom-up, stimuli-driven: a Cognitive Agent has its internal drives. I.e. the
axiological and operational counterpart is driven by the motivational apparatus, which embodies the
desires of the system that compete in order to activate their functions.
In a Cognitive Agent, descriptions embody a point of view that allows to constrain perceptions and
can be applied to new problems; seeing-as is the natural modality of categorizing and addressing
problems, because solutions can be extended to all the problems that share some common structure,
despite of possible differences in their content.
Problem solution schemes may be used as a problem position language to conceive and pose new
problems. This even means that having a language that allows us to pose problems makes the space
of possibly-met problems much larger: imagination can substitute experience.