It is our hope that Principia Cybernetica's epistemology and ontology based on cybernetics
are not only of purely philosophical interest, but may be used
in solving an important problem which stands wide open at the
present time: creation of a Universal Semantic Language to be used
in the emerging Global Brain.
We can distinguish two levels of cybernetic systems involved in
the global brain:
- individual human brains, and
- computers and other cybernetic devices united into one huge system, such as the Web
.
Individual human brains exchange information by using various languages,
of which the most important are universal natural languages,
such as English or Russian. Computers exchange information by using
formal computer languages of various kinds, which can be considered
as one universal computer language. There is a gap between these two
levels, and this constitutes a problem. We understand computer languages,
but computers do not understand our languages: they are informal,
the meaning of linguistic objects is inseparable from the human brain.
What we need is a formal universal language to catch the meanings
present in natural human languages in a form understandable by
computers.
This problem is well known. The reason why it is not yet solved
is simply that the problem is immensely difficult.
Practical people get satisfied with partial solutions:
creation of narrowly specialized sublanguages which superficially look
like human languages. Financing of more ambitious projects is tight,
because nobody can promise fast returns on this path.
Why is our philosophy relevant?
Developing a universal semantic language we must start with some
elements which are absolutely minimal and primitive, so that
all representations that are involved in human laguages could be
expressed in terms of our semantic language as constructions from
those primitive elements. Otherwise it is hardly possible to
obtain the formality that is demanded by the computer.
Our ontology states that such primitive elements are actions,
and nothing but actions, while objects are relation between actions.
Our epistemology states that the meaning
of linguistic statements is in the predictions they produce.
But the concept of a prediction takes us, again, to the action
whose termination it states. The other kind of linguistic objects,
commands return us even more directly to actions.
Thus we have a closed basis for a formal system.
So, why not try? Any ideas and accounts of relevant work are welcome.
Representation of semantics on the basis of formal logic and
some primitive predicates, which is used in much work on artificial
intelligence, is, certainly, relevant. Our goal should be going
further down in analysis than it is usually done -- a goal that
requires a consistent and all-embracing philosophy.
