What we are talking about is the time it takes for a mass of particles, the condition of the
universe shortly after the big bang, to evolve into a set of highly complex astronomic structures.
It is not difficult to see that the answer cannot be found through pure probability. However,
the time problem can be overcome if everything in the universe was composed of layers of
simpler things. Simon explained this in a very famous parable. The parable was repeated in a
number of works, this version was taken from Howard Pattee's Hierarchy Theory, The Challenge of Complex Systems.
One can show on quite simple and general grounds that the time required for a complex
system, containing k elementary components, say, to evolve by processes of natural
selection from those components is very much shorter if the system is itself comprised of
one or more layers of stable component subsystems than if its elementary parts are its
only stable components. The mathematics of the matter is a straight-forward exercise in
probabilities. But the gist of it can be given even more simply in a parable.
Two watchmakers assemble fine watches, each watch containing ten thousand parts.
Each watchmaker is interrupted frequently to answer the phone. The first has organized
his total assembly operation into a sequence of subassemblies; each subassembly is a
stable arrangement of 100 elements, and each watch, a stable arrangement of 100
subassemblies. The second watchmaker has developed no such organization. The average
interval between phone interruptions is a time long enough to assemble about 150
elements. An interruption causes any set of elements that does not yet form a stable
system to fall apart completely. By the time he answered about eleven phone calls, the
first watchmaker will usually have finished assembling a watch. The second watch-
maker will almost never succeed in assembling one--he will suffer the fate of Sisyphus:
As often as he rolls the rock up the hill, it will roll down again.
It has been argued on information-theoretic grounds--or, what amounts to the same
thing, on thermodynamic grounds--that organisms are highly improbable arrangements
of matter; so improbable, in fact, that there has hardly been time enough, since the
Earth's creation. for them to evolve. The calculation on which this argument is based
does not take account of the hierarchic arrangement of stable sub-assemblies in the
organisms that have actually evolved. It has erroneously used the analogy of the second.
unsuccessful watchmaker; and when the first watchmaker is substituted for him, the
times required are reduced to much more plausible magnitudes. Specifically, on the
simplest assumptions, the mathematical model shows that if a system of k elementary
components is built up in a many-level hierarchy, and s components, on the average,
combine at any level into a component at the next higher level, then the expected time
of evolution for the whole system will be proportional to the logarithm to base s of k. In
such a hierarchy, the time required for systems containing, say, 1025 atoms to evolve
from systems containing 1023 atoms would be the same as the time required for systems
containing 103 atoms to evolve from systems containing 10 atoms. The form of the
generalization is interesting, in that it describes a relation between two levels of a system
that is independent of absolute level.
We conclude that hierarchies will evolve much more rap-idly from elementary
constituents than will non-hierarchic systems containing the same number of elements.
Hence, almost all the very large systems will have hierarchic organization. And this is
what we do. in fact, observe in nature.
What does this approach tell us. First, we began with an assumption, that the world is rational.
This is not knowledge. We accept it only on faith. But this faith has been buttressed by the
simple fact that we have been acting on this assumption for 2600 years and we haven't gotten
into too much trouble from it. This is true even though if an event happened that violated that
assumption we would either immediately supply an ad-hoc reason for it or simply ignore it.
But the same is true of experience. We train ourselves to identify objects in our environment.
One characteristic of connectionist neural structures is that they have the ability to recognize
things even though their presentation is highly distorted. That is why these are capable of
doing things like reading hand-written zip codes. As a result they make mistakes. You see
someone in the distance and identify him as your friend. As she comes closer you discover that
it was not him after all. You may even wonder how you could have possibly mistaken her.
That is the nature of these structures. They are valuable because they are right most of the
time. This is the real meaning of the coherence theory of truth. It means that the world makes
better sense if we accept it as true enough.
Simon stated what seems to be an obvious facet of the world around us, that everything in it
appears to built out of a hierarchy of smaller things. This, of course is what we call an
empirical generalization. Again we accept it as true enough because our limited experience
seems to verify it and we don't get into any trouble if we act as though it were true. We are
doing pretty well so far, can we find any more of these kinds of "facts" that seem to be the
way the world around us is structured?
The properties of anything are not related to the properties of what it is made up of. Does this
statement fall into the category of a "fact" that is "true enough"? If it were not true the
properties of hydrogen and the properties of plutonium would be alike. They both would be
related to the properties of the sub-atomic particles that make them up since the sub-atomic
particles of one are completely interchangeable with those of the other. Oh, you say, your
wrong. The properties of hydrogen and plutonium would be completely different if they were
composed of oranges and chestnuts instead of electrons and protons. And you would be right.
Partly. This is what Simon called loose decoupling. Beyond those characteristics that are true
of everything that is composed of sub-atomic particles of the nature of protons and electrons,
the properties of all of the elements are determined by the organization of particles and not the
particles themselves. But these properties that are common to all elements are determined by
their being composed of sub-atomic particles and not by the properties of the particles
True? Well, maybe good enough. You see the property of "loose decoupling" means that the
properties of the elements are not bound by the properties of the subatomic particles. Not only
this, but we can generalize from this to the idea that every complex system derives it
predominant properties from the organization of the system and not from the properties of its
Are these statements true? Everything that we experience in the normal run of our lives is
composed of something smaller, faster, or less complex. We have faith that all, or at least most
of what we have not experienced is probably organized similarly out of other components. Our
faith is buttressed again by the fact that we have acted on the basis that these things are true
for a very long time and they haven't gotten us into trouble. This is what John Dewey meant
when he said that truth is a warrantable assertion.
If this is true, then. If the predominant properties of something that exists are determined not
be the properties of its constituents, but by the organization, then there is no way we can
determine the properties of the elements of such a system from the properties of the system.
We must first isolate the elements and then determine their properties. Therefore, why must we
assume that something that we know exists, that is something we know the properties of, must
be composed of physical entities? Let us begin with connectionist structures. If you were to
examine the brain even with the most powerful instruments you could never find these
structures. They are organized interconnections between neural nodes. The interconnected
nodes need not be physically adjacent. They need only be connected to one another. In fact
there is a good reason why it is better if they are not adjacent. Damage to the brain would
have to be much more extensive for it to destroy a connectionist structure if it is distributed
throughout the brain rather than located in one spot. The purpose of these structures is very
much like what Aristotle considered imagination to be. They register the existence of objects.
In this case objects they are trained to recognize. The inputs to the structures are signals from
the various senses in the body. The output is information.
The output from a connectionist neural structure is a pattern of output states on the output
layer of the neural structure. Remember, the information is coded in the pattern of the states,
not the states themselves.. The pattern that is formed by all of these structures that are active
in the brain at any definite time is a higher level information structure. This pattern is not
made up of states, it is made up of patterns of states. This pattern, then, is our awareness of
the immediate environment around us. We are now well beyond anything that we can call
physical. Yet we still have not located the mind.
To meet the demands of the metaphysicians mind must be disconnected from the body. The
difference between the entity we call mind and those structures that we have been discussing up
to now is that these are created through an interaction between the senses, the neural networks
in the brain, and the outside world. The mind, because it must create out the raw material
delivered by these systems a living world in four dimensions, must be creative rather than
created. Consider the computer I am typing on right now. There is a physical structure
involved, the microprocessor, the memory, the magnetic storage devices, and the peripherals
that make the system work. Then there is the operating system. This is a set of coded
instructions that act as a translator between the program and the hardware. The program too
exists as a set of instructions that convert what my activities into procedures for the operating
system to follow. But I am not running a program. I am typing letters on a screen. As soon as
the program begins to run it is more than a set of instructions. It is an interactive media
responding immediately to my actions. The properties exhibited by the process have very little
to do with a set of instructions and nothing to do with the states of electronic gates in the
computer itself. So we are used to dealing with things that are of a sufficiently high level of
complexity that we can treat them as though they were things in themselves rather than the
actions of physical entities like electronic gates.
The modern computer is not even a close approximation to the neural network that constitutes
the human brain. The human mind is the result of an extremely high degree of complexity. To
say that it is nothing but the interaction between neurons is the same as saying that this book,
at least until I type it out, is nothing but the states of a lot of electronic gates.
The heart of the computer I am using is the microprocessor. It is in itself a highly complex
device. But is made up of a large number of very simple elements. It is the organization of the
elements and it is the organization of the procedures through which the microprocessor operates
that makes it such a powerful tool. The mind cannot be traced down to neural networks any
more than this book can be traced down to patterns of electronic gates.
You can not shut off a mind. When the electrical activity in the brain ceases the person is dead
and cannot be revived because the mind does not exist as a pattern of static nodes, it is a highly
complex pattern of neural activity. Thus it is not physical like the connectionist structures in
the brain even if it depends for its existence on physical entities like networks of neural nodes.
This is important for more reasons than the solution of a long-standing philosophical problem.
It is important for our view of the future of Western, or more likely soon to be Terran culture.
The World Wide Web is a structure of interconnected communication networks. We can identify
the computers, the telephone wires and optical cables, the physical hardware that make up the
net. But as the net evolves these become irrelevant. Technicians can change out a whole bank
of computers for an entirely different kind, change from copper wires to fiberglass cables, even
change the protocol by which the various nodes communicate and those who surf the net would
never be aware that a change took place. We can no longer think of things that exist as
physical entities. The next Descartes will open up ways of understanding our world that are as
revolutionary to us today as those of his seventeenth century predecessor's ideas were to his
contemporaries. We are living through the most exciting period in the history of the world. I
have absolutely no doubt that some man will rise to the challenge. I guess it is my act of faith.