Globalization of Knowledge and Insight
Envisaging a paradigm-shifting software package
- / -
The concern here is the possibility that, with relatively little cost and effort, a
software package could be designed and disseminated that would make a major difference to
the way in which individuals and groups organize the knowledge of concern to them.
The specific concern here is to highlight 'global' dimensions in the
organization of knowledge. Global is taken here to mean integrative in ways beyond those
currently achieved by setting up knowledge hierarchies and checklists of topics. Global is
not taken to mean world-wide in the geographical sense that is now characteristic of most
discussion of 'globalization'. Rather global is conceived as integratively
meaningful to a person or a group. It is also to be understood as holding together the
concerns of a policy-maker faced with the full messy array of societal issues. The
challenge of interrelating diverse perspectives, including those from a variety of
cultural perspectives, is also inherent in this proposed understanding of global.
The approach taken here is necessarily relatively simple, if not simplistic. However
the software framework envisaged is designed to be developed and to encourage further
development by users and those with software skills. It is not a closed and completed
system. This paper builds on arguments developed in earlier papers by the author on this
theme (Judge, 1994-1997).
Principal themes: a point of departure
There is no lack of software to handle nested hierarchies of topics. There is very
little that addresses the challenge of how the principal themes are organized in relation
to one another -- other than as checklists of general topics (agriculture, sport, etc).
But the agonizing challenges of society, especially for policy-makers, concern precisely
how the most general themes relate to each other. It is at this level that the need for
distinct government ministries is decided. It is at this level that the necessary
communications between such agencies become apparent. This is also true in relation to
local government departments. It is also true for the individual in allocating resources
between disparate activities (leisure, food, transportation, religion, etc).
The first difficulty is that there seems to be relatively little understanding of what
are the principal themes most closely related to any particular theme. For example, if
'sport' is taken as a principal theme, what are the most closely related or
relevant principal themes -- where the meaning of such proximity remains to be more
precisely defined? In identify such proximates, there is also the need to clarify whether
they are the 2 closest, the 3 closest, or the N closest. The resulting picture might
- Health, Leisure / Leisure, Media / Leisure, Education
- Health, Leisure, Media / Leisure, Media, Education
- Health, Leisure, Media, Education
In this case, at Level-2, three sets of 2 proximates have been provisionally
identified. At Level-4, for example, only one set of 4 proximates is presented above.
Clearly views will differ on what are the principal themes (tops of hierarchies). It
should be up to the user to extend the choices at each level and to give preference to
choices at any one level -- and indeed to prefer a particular level (such as the single
choice presented at Level-4). The software could hold a library of possible choices for
proximates for principal themes. The user could add to this library, or could save
The next concern would be the number of principal themes to be explored in this way.
Past work in classification by Ingetraut Dahlberg identified 100 themes presented
in her International Coding Classification (1982). This was considerably adapted
by the present author (Judge, 1983) as the basis for the classification system
used by the Union of International Associations in both its Yearbook of International Organizations
and its Encyclopedia of
World Problems and Human Potential. The number of principal themes remains
100 (as illustrated in the web demonstration of this approach at http://www.uia.org/topics/aaintmat.htm).
Although covering very different spreads of subjects, the number of ministries
in any one national government seldom exceeds 100.
The concrete example of government ministries illustrates the fact that 'principal
themes' do not need to be single word themes. For example, in the case of a Ministry
of Education and Culture, the principal theme might be 'education and culture'.
The question of why they are so closely associated and whether they are distinguishable
would need to be reviewed by the government concerned. If however, within that ministry,
there were two major departments, 'education' and 'culture', there
would be a case for treating these independently in the framework outlined above. The
software package is designed to respond to the perceptions of the user, not as an
imposition of a predetermined set of categories.
The result of the exercise would therefore be a database with some 100 or more records
-- one for each principal theme as perceived by the user. There is no restriction on the
number. A user would be free to include many more such themes, or far fewer. The user
would be free to accept (amend, or reject) any 'standard', or recommended, set
of themes emerging from the information science community.
There is also merit in offering the user the possibility of forcing in (or excluding)
an open category 'Unforeseen', which could serve as a kind of 'joker'
in the set. Including this would recognize the possibility of surprises and emergent
themes. Excluding this would signal an assumption that the system of themes was considered
complete and closed.
There are many symmetric and asymmetric patterns -- many known only to those with
mathematical interests. With the emphasis of this package on 'globalization of
knowledge', the following discussion relates primarily to centro-symmetric patterns
in three dimensions. The case of two dimensional tiling patterns is excluded, although it
merits exploration as an option in this context. Interesting forms like the torus are also
Essentially the focus here is on the tiling patterns possible on the surface of a
sphere -- as the simplest representation of globality in three dimensions. The concern is
how to 'fit' together tiles, where each tile represents a principal theme
defined by the above database -- and given that each record in the database indicates what
are the preferred neighbouring tiles according to the user's perception.
The assumption here is that some form of paradigm shift occurs when knowledge can be
successfully represented globally by patterning it over the surface of a sphere. In
contrast with conventional checklists of subjects, presentation in this new way amounts to
a form of 'globalization of knowledge'. Just how this might be more than a
gimmick, will be explored later.
The software therefore needs access to a library of patterns. These patterns could be
indexed by the number of their faces, edges and apices. These are the conventional
attributes of centro-symmetric polyhedra.
Consider a case where a user had defined 72 principal themes. The software needs to
access a pattern with 72 faces. There may be several possibilities.
Having accessed an appropriate pattern, the software then needs to use the information
in each database record to endeavour to attribute each theme to a particular face -- such
that it is contiguous with other faces representing themes perceived as proximates in that
database record. Each face thus represents a thematic area.
Respecting the user's specifications, there may not be a neat solution. The software
should then consider and propose alternatives. This could be done with more or less input
from the user:
- Software initiated solutions: The software could for example propose a pattern
with more faces, some of which were unattributed, in order to fit all the themes
appropriately onto the surface.
- User initiated solutions: The user could tag the database records with harder and
softer preferences. This would then signal the software where the user absolutely required
a fit, and where the software could override the user's selection. For example, the user
could indicate for sport that the Level-4 pattern was a 65% preference, but that the
Level-3 pattern was an 85% preference. In this case the software would seek solutions with
the Level-4 preference, but would fall back on the Level-3 preference if these could not
After endeavouring a match, the software could indicate where changes of preference
would enable a better match to be made.
Presentation of the pattern
Once a solution has been produced by the previous process, the pattern could be used
with the thematic labels to generate a three dimensional structure using the virtual
reality markup language (VRML). The faces could be appropriately coloured and labelled.
Hyperlinks could be attached to each faces so that when the user clicked on them, this
would result in access to a distinct web page -- or to another VRML structure.
Using a virtual reality browser (now available free of charge), the user then has,
literally under finger tip control, a global representation of the pattern of knowledge
implicit in the original database. This display can be rotated, and explored by zooming --
with access at any point to underlying information. The spherical representation is a
globalization of the user's knowledge world -- previously only accessible through
checklists, or possibly mind-maps. An illustration of some of these possibilities is
available from the web page: http://www.uia.org/uiademo/vrml/vrmldemo.htm)
Changing assumptions: Initially it is to be expected that the prime function of
such a display would be to raise questions as to its adequacy -- notably with respect to
the user's assumptions in the database from which it was generated. It would lead the user
to experiment with alternatives, possibly by changing the weightings attached to items at
different levels for a given theme. It might also lead to the splitting or merging of
Incompletion: The user might be especially interested in patterns with
'incomplete' solutions, namely those in which specified contiguity could only be
achieved by having one or more unattributed faces -- 'oceans' separating
'continents' of contiguous 'countries', possibly with
'peninsulas'. 'Islands' might also reflect the user's reality in some
cases. Alternatively each separate 'continent' could be generated as a distinct
global representation. The user's knowledge system could then take the form of separate
'planets' revolving around the user's central point of reference.
Sub-themes: The user might also want to explore the sub-themes within any given
theme. In the case of a given government ministry, for example, it would be composed of
many departments with particular thematic responsibilities. A separate database could be
set up, with one record for each such sub-theme -- along the same lines as the initial
database. The pattern matching procedure could then be repeated, resulting in the
generation of a sub-theme global representation. Using the hyperlink facility, clicking on
a thematic face of the original representation could transfer the user to the
corresponding sub-theme representation. In this way virtual reality techniques can be used
to hold many levels of knowledge. These sub-worlds could also be held as 'moons'
around the parent world, although this can clutter and confuse the overview.
Imagery: Virtual reality techniques also permit (coloured) images to be attached
to each of the faces, instead of (or in addition to) simple text labels. The user could
then specify mnemonically significant images (as image filenames) in the original database
for use by the software in building the final global representation. It would be expected
that the user would continue to tinker with such colour and image attributes. In the case
of the United Nations system, for example, the images might be the logos of the various
Directory listings: Clearly the global representation serves, at the highest
level, as an integrative substitute for such tools as the 'explorer' or
'file manager' in Windows software. Ideally it could also enable the user to
access subdirectory structures in the conventional manner.
Global patterning: Beyond the facilities indicated above, there remain the
possibility that the pattern of faces may be used to carry other levels of information.
Typically, for example, clusters of faces on the surface of a centro-symmetric polyhedra
themselves form sub-patterns. Sequences of edges form circular routes around the sphere --
whether around the sphere as a whole, or around regions of it..
Significant relationships: These patterns merit considerable attention. They
suggest the possibility of significant, or preferential, lines of communication between
certain thematic areas. In the case of government ministries, they raise the question of
who might usefully be communicating with whom. They may suggest useful feedback loops that
Relating incompatible perspectives: Possibly of greater interest are those
thematic areas that are effectively distant from each other. Manipulating the global
representation using a virtual reality browser, these are the areas which may not even be
visible simultaneously -- it may be necessary to rotate the structure to bring a distant
surface into view. This operation reinforces understanding that a global representation of
knowledge may only be possible if certain perspectives are effectively invisible to each
other -- possibly reflecting quite incompatible perspectives in practice. They are beyond
each other's 'horizon'. The global representation is then effectively a coherent
framework that 'holds' such incompatible perspectives.
Emptiness: This whole approach suggests that the 'globalization of
knowledge' may only be possible around what amounts to a central, unoccupied
'emptiness'. It is structuring with respect to this emptiness that allows the
disparate themes to be meaningfully ordered. The integrative dimension is intimately
associated with the structural role of the unoccupied area in the representation.
As a software project, what is proposed is feasible with limited resources. In its
basic form it would allow many users to offer themselves an overview of the knowledge
context within which they work. It has the merit of allowing users to adapt it to their
own needs and to experiment with it in unforeseen ways.
In the process of producing the package, refinements should quickly become apparent
with the assistance of mathematicians. These should make it easier to ensure that projects
are 'globalized' rather than undertaken as 'flatland' initiatives. For
example, 'universal declarations' might usefully position the themes of their
principal clauses as faces on such a global representation. This should highlight the
complementarities between different principles as the basis for a coherent whole. It
should prove helpful in designing lines of communication appropriate to any explicit
pattern of checks and balances in an institutional or community system.
Ingetraut Dahlberg. ICC -- Information Coding Classification: principles, structure
and application possibilities. International Classification, 9, 1982, 2, pp
- Functional classification: a review of possibilities. In: Global Action
Networks. K G Saur, 1983, pp. 737-755
(Amended annually as a supplement to Volume 3 of the Yearbook of International
- Devising a paradigm-shifting device. 1997 [text]
- Sustaining the coherence of dialogue through apartness; patterns of
systematic configuration of entities through hypertext. 1997[text]
- Envisaging the art of navigating conceptual complexity 1995 [text]
- Spherical configuration of categories 1994[text]
- Strategic correspondences: computer-aided insight scaffolding 1995
Union of International Associations. Yearbook
of International Organizations. Munich, Saur, 34th ed, 1997/98,
Union of International Associations. Encyclopedia
of World Problems and Human Potential. Munich, Saur, 4th ed, 1994-5 [commentary]