E Pluribus Unum - Computational Creativity Research: Towards Creative Machines - page 180

two bases by a bridge theory. For example, the bridge might introduce a higher-level
connection between two image schemas which is then used to create the blended
theory. An example of this might be where we have image schemas involved with
time and with space and combine these first for the definition of a theory in which
time and space are linked (as in our navigation examples above or even in the linking
between time and space discussed for comics and visual narrative).
Technically, such diagrams are closely related to alignments [ 9 , 74 ], and to distrib-
utedmodelling languages [ 40 ]. Concerning the formalisation of conceptual blending,
these techniques and diagram patterns will be of particular importance to tackle a
computational treatment and formal representation of so-called generic integration
templates (GIT), i.e. the idea of studying general blending templates, first introduced
and discussed in detail by Pagán Cánovas [ 59 ], with more examples to be found in
Turner [ 68 ].

9.5 Blending in the Hub

In this section we will discuss the computational and representational support for
formalised blending via the Ontohub.org platform as well as the evaluation problem.

9.5.1 Representation and Computation

To begin, combinations (or, alternatively, the underlying colimits) can be computed
directly by the web platform Ontohub. Ontohub is a repository engine for managing
distributed heterogeneous ontologies. Ontohub supports a wide range of formal logi-
cal and ontology languages and allows for complex inter-theory (concept) mappings
and relationships with formal semantics, as well as ontology alignments and blend-
ing. Ontohub understands various input languages, among them and DOL .

We describe the basic design and features of Ontohub in general, and outline
the extended feature-set that we pursue to add to Ontohub for conceptportal.org—
a specialised repository for blending experiments within the distributed Ontohub
architecture.
The back-end of Ontohub is the Heterogeneous Tool Set

OWL

, which is used by
Ontohub for parsing, static analysis and proof management of ontologies.

Hets

Hets can

also compute colimits of both and first-order logic diagrams and even approx-

imations of colimits in the case where the input ontologies live in different ontology
languages [ 8 ].
Computation of colimits in

OWL

' general colimit algorithm
for diagrams of sets and functions (note that signatures in most cases are structured
sets, and signature morphisms structure-preserving functions). Such a colimit of sets
and functions is computed by taking the disjoint union of all sets, and quotienting
it by the equivalence relation generated by the diagram, which more precisely is

Hets is based on Hets