The A3 approach proceeds from a vision on the process of formation of mind content,
treated as a network consisting of mental entities. Also, the formation of this network is
treated as „weaving“ rather than „creation of a whole“ because this process is perpetual and
it never ends during the lifetime of mind. Since the term „formation“ is too general to reflect
the specifics of this process, and „weaving“ is not a convenient word to serve as a term, I
will use the term „synthesis“ for such formation of mind content. Currently, there exist a
large number of different types of networks used in research and technology – neural
networks, semantic networks, Petri nets, and other types. Mind content is a new type of
network.
Synthesis unfolds in steps and for a natural (i.e. not „artificial“) mind, probably, a large
number of mental entities is created at one step. In order to simplify the analysis, in this
paper I will limit to only one atomic mental entity created at one step of the mind content
synthesis.
Since the mind content synthesis is regarded as perpetual and never ending during the
lifetime of mind, it always encreases in size and, even though the mental entities are more
abstract than sensual data, the mind content is always a realistic „film of the reality“ - a
„conceptual film“. I am not interested here in the dynamic aspect of mind content which
determined me to assimilate it with a „conceptual film“. The dynamic aspect only shows
that mind content is an „open world“, i.e. a changing object. I am interested in mind content
as a „history“ or „memory“ of the synthesis process reflecting the „structure“ or „form“ of
the mind content. The Figure 1 below illustrates the „mathematics of brain“ and serves as a
|
|
roadmap to further material.
58 Web Intelligence and Intelligent Agents
Fig. 1. Brain Mathematics
Earlier (Drugus, 2007), I treated the synthesis of mind content without the part-whole
aspect. I explained that such synthesis is done by alternative involvement of each of the two
hemispheres of brain at one step of synthesis, and with each involvement, it does exactly
one operation whereby it creates an entity. For right-handed people, I substanciated that the
left hemisphere is specialized on the association operation whereby it produces a new entity
also called association, while the right hemisphere is specialized on the aggregation operation
whereby it produces a new entity also called aggregation. In this paper, I will additionally
treat the part-whole aspect, which became possible due to finding a better explaination of
the part-whole aspect, dealt with by the third component of brain – the bridge between two
hemispheres called corpus callosum. I assume this component to be specialized on the
operation of atomification of the A3 approach.
The mind content obtained in result of a perpetual synthesis can serve only as a basis for
intelligence, which alongside accumulation of conceptual experience as mind content,
manifests via other activities, including the logical activity. But the mind content synthesis
unfolds on a level which serves as a basis for higher levels of intelligence. The level
immediately above it allows to cut on this wealth of conceptual data and can be said to be
the level of identification. The identification allows to regard many mental entities as „the
same“. Because these two layers are sufficient for the discipline of Universics, in this paper,
I will focus only on the level of mind content synthesis and on the level of identification.
To summarize the account of the A3 approach vision on brain mathematics, I will say, that
at one step of mind content synthesis, the brain applies exactly one of the three operations,
said to be A3 operations, for creating a new entity:
- Aggregation operation for creating an entity also called aggregation – an entity type
generalizing the notion of set of set theory, and the notions of class, container and
collection of Semantic Web,
- Association operation for creating an entity also called association – an entity type
generalizing the notion of ordered pair of set theory, and the notion of property of
Semantic Web,
- Atomification operation for creating an entity also called atomification – an entity
type generalizing the notion of atom (also said to be ur-element) of set theory and
Universics - a Common Formalization Framework for Brain and Web 59
the notion of ‚whole’ of mereology, science of „parthood“ or of „part-whole“
relationship (and for which there is no corresponding notion in Semantic Web).
|
|
To clarify the above, it should be added that the entities are either untyped or typed, and if an
entity is typed, then it can have one of the three types above, which I will refer to as the A3
types. Each of the terms aggregation, association, atomification, will serve as a name both for the
operation and for its result, and we will distinguish between them by specifying to which of
the two we refer. So, we will use expressions like „an aggregation“, „entity of type
aggregation “, or „aggregation operation“ and similar expressions for „association“ and
„atomification“.
In the most general case, I am treating the term operation as a many-to-many
correspondence. So, even though A3 operations are not yet completely defined, they should
be treated as many-to-many correspondences or, how they are called elsewhere – „manyvalued
functions“. The notion of property in Semantic Web corresponds to the notion of set
theoretic relationship and of „functional property“ as it is said to be there. This is why to
avoid term collision, I will not say a many-to-many correspondence to be a function - I will
say it to be operation.
One of the greatest difficulties in guessing how the mind content is synthesized is to find
out how the representations of different objects can be retrieved at a later time. This is a
general problem for any kind of networks. Also in philosophy the process of discriminating
separate entities in the „material network“ of the Universe has been a problem since ancient
times – a problem, which has not been solved up to date.
I would have preferred the term „domain of integrity“ for the representation of an object in
the network of mind content, because this would have better reflected the idea of dealing
with the „wholeness“ treated here as a synonym for „integrity“. But „integrity“ has a special
meaning in English, and I will use the term „integral domain“, which though is good due to
being shorter. I consider the integral domains, as representations of objects, to appear in
mind due to the atomification operation.
Fig. 2. A piece of mind content
The figure 2 above illustrates the mind content obtained in result of a synthesis process,
which is a directed process: in this figure, we consider that mind content „grows“ upwards.
The results of applications of both the aggregation and association operation are represented
as nodes of the network. In order to simplify the picture, the distinction between the entity
types aggregation and association is not indicated in the figure and all the aggregations in
60 Web Intelligence and Intelligent Agents
this figure are deemed to be two-element aggregations – same number of arguments as of
the association operation. The colored areas are said to be integral domains. These pieces of
mind content are „production“ of the atomification operation, which can be thought to act in
a dimension, different from the dimensions where act association and aggregation. The
results of applications of the atomification operation can be imagined to lie in space above
the 2D space of the picture. This is why, to represent the integral domains as new entities
not lying in the plane and to be able to discriminate between them, they are represented in
different colors. The bottom of each integral domain is represented by the atoms (indivisible
constituents) of the represented object. This botom is „dented“ because in the synthesis
process the representations of the atoms appear at different times. The top of each integral
domain is represented by exactly one entity which we regard as the unity of this integral
domain. The first intuitive idea about the manner how an integral domain can be extracted
from the mind content is by imagining that such a domain is “carved out” from the mind
content taking into account the two boundaries - top boundary represented only by the
unity, and bottom boundary represented by many atoms.
Notice that in the Universe of matter, one body B is part of another body C, if B is spatially
“inside” C. The universe of mind differs from the Universe of matter and the
representations of B and C in mind look totally different. Namely, the unity of B becomes an
atom of C, so that in a graphical representation as that of Figure 2, the representation of B
would be beneath (and not inside) the representation of C.