core role that “self” plays in the generation of sophisticated behavior is undeni-
able. Indeed, part of the importance of self-consciousness is related to distinguish-
ing oneself from the emvironment in this class of models (e.g. for action/agency
attribution in critical, bootstrapping learning processes).
7 Stepping to Awareness and Consciousness
Our strategy for consciousness research is not following Alexander ’s approach of
axiomatising consciousness (i.e. searching for a complex predicate to be accepted by
the community) but of analysing and formalising the core issues and mechanisms
involved (and addressed in the principles on cognitive control exposed so far). The
reason is simple; while the axiomatisation process is important for clarifying the is-
sues, it may give rise to a receding horizon phenomenon similar to what happened
to AI in the eighties. This not happening, both approaches should lead to the very
same end.
7.1 Defining awareness
From the analysis of integrated cognitive controllers given in the previous sections
we can make a try into the formalisation of some consciousness aspects. We will
make a distinction between awareness and consciousness, reserving the C-word
for systems self-awareness.
Principle 6: System awareness — A system is aware if it is continuously
perceiving and generating meaning from the countinuously updated models.
The term meaning was introduced in this principle to define awareness and this
looks-like eluding the core definitional problem. However, the word meaning im-
plies that the main difference between perception and awareness is the addition to
the perceptual mechanics of a certain value system in the global system process. So
we can say that awareness implies the perception of value to the system from its
sensory flow.
The updated integrated model produced by perception is evaluated in terms of
a value system not only in the present state of affairs but in the potential —future
and past10— consecuences derived from this state of affairs. Awareness implies the
partitioning of predicted futures and postdicted pasts by a value function. This par-
titioning we call meaning of the update to the model. In this context of interpretation
of the term meaning, we conclude that only pieces of information that are model-
integrable can have meaning, because for others, we cannot compute futures nor
pasts, less their value.
10Restorpective prophecy in the words of T.H. Huxley Huxley (1880).
ASLab.org / Principles for Consciousness / A-2007-011 v 1.0 Final
14