2

THE CLOCKWORK
UNIVERSE IN CHAOS
"How could they formulate laws
that would curtail the freedom of God
to change His mind? Nevertheless
they did, and they got away with it"
—Stephen Hawking

"The often repeated statement,

that given the initial conditions we know
what a deterministic system will do
far into the future, is false."
—Predrag Cvitanovic

O n a Caribbean island, a butterfly happens to flutter briefly to the

left instead of to the right. As a result, minute swirls of air pro-
duced by its wings move in a slightly different way than they would
have. A few days later a hurricane that had been gradually building up
force veers into the Florida coast instead of heading out to sea.
Could the hurricane's change in course have been caused by the
altered flight of the butterfly? According to Edward N. Lorenz of the
Massachusetts Institute of Technology, this is actually possible (Lorenz,
1963). Computer simulations carried out by Lorenz suggest that the
flow of air in the atmosphere may display the property of "exponential
instability." Under such conditions, arbitrarily small changes in the
flow pattern will be amplified until they have a major impact on the
turbulent pattern of flowing air that makes up our weather. According
to Lorenz, this makes short range weather forecasting extremely un-
predictable, because of the impossibility of monitoring the very small
changes that could possibly result in large-scale effects.

29
 This unpredictability has been called "deterministic chaos," since
it arises in systems that should be strictly deterministic and predictable
from a mathematical point of view. The concept of deterministic chaos
sheds significant light on the question of whether or not the mind can
con l rol the brain without violating the laws of physics. It also sheds light
on I he related question of how God might control the processes of
nature. Put simply, if undetectable changes can control the unfolding
of big effects, then nature can be intelligently controlled without
measurable violation of the laws of physics.
I Infortunately, this idea is unsatisfactory as it stands, because even
it deterministic chaos allows for mind-brain or God-world interaction,
this only tells the physical side of the story. We still don't know what
ca uses the undetectable, guiding changes. However, we can model such
causes by taking advantage of the idea of virtual reality. Suppose, for
the sake of argument, that deterministic chaos provides a way to pro-
gram a virtual reality system in which a human user interacts with a
physically lawful virtual world. We can then argue that something simi-
lar could apply to the Ground Reality model—the hypothetical virtual
reality system of the real world.

NEWTONIAN PRECISION
I shall outline how such a virtual reality could be constructed. But
to lay the groundwork for this, I first go back to the 17th century, the
e i a of Rene Descartes and Isaac Newton. At this time, educated people
in Murope widely accepted that the material world was both created
and directly controlled by God. However, Newton's Principia gave
si i ength to the seemingly contradictory view that everything in nature
happens in a rigidly deterministic way under the control of fixed
mathematical laws.
Newton's work seemed to confirm an image of reality that had
been growing more and more prominent in Europe since the late
M idtlle Ages—the picture of the universe as a machine comparable to
a vast clock. According to this understanding, there are two ways in
which God can influence the behavior of matter. These are (1) by
setting upthe world-clock and letting it run by itself and (2) by changing
the clock's settings from time to time. Here God may be thought of as
 31

transcendental (lying beyond the material realm), or as immanent (per-

vading matter). Either way, the predominance of clocklike laws of
nature forces God into the role of a maker or adjuster of clockwork—
an idea that did not arise in earlier conceptions of nature as an organic
whole.
Newton himself felt that the solar system was constructed by divine
creation and that intervention of type (2) was necessary for its contin-
ued smooth operation. However, Newton's theistic arguments were
far from acceptable to many scientists and philosophers of his time.
Newton's rival, Leibniz, for example, preferred to think of God as a
perfect clockmaker who, after creating the universal machine and
setting it in motion, had no need to intervene further in its operation.
In due course, this speculative theological argument was but-
tressed by the theoretical researches of Pierre Simon de Laplace and
J. L. Lagrange concerning the stability of the solar system. Newton had
argued that the mutual gravitational influence of the planets on one
another would eventually perturb their orbits to a degree demanding
divine correction. However, Laplace and Lagrange showed by math-
ematical analysis that in an idealized model of the Newtonian planetary
system, these perturbations would cycle periodically within fixed lim-
its. (Recently, however, orbital studies aided by computers have called
this into question and have suggested that the solar system is chaotic
and unpredictable in the long run.)
As time went on, the idea of divine creation also began to seem
more and more implausible. Newton had argued that the regular
arrangement of the nearly circular planetary orbits required the "di-
vine arm" (Newton, 1692-3). But Laplace suggested that this regular
pattern may have formed naturally as the planets condensed by physi-
cal processes from a primordial nebula. Even though this hypothesis
was not worked out mathematically in any detail, it carried the day,
and further limited the scope allowed for God's activity within the
universe.
These theological developments were paralleled by the develop-
ment of scientific views regarding the mind and the body. It is no
coincidence that Descartes' model of the mind and the brain was pro-
posed during a time when mechanistic models of reality were becoming
more and more prominent. As physical laws became deterministic and
32

clocklike, it became difficult to see how a nonphysical entity could

influence the behavior of matter. Yet, if such an entity did not influ-
ence matter, then it was automatically "kicked upstairs" to an irrel-
evant state of pure transcendence that is tantamount to nonexistence.

CHAOS AND DETERMINISM

But the problem may lie more with the metaphor of a clock than it
does with the laws of physics themselves. To see this, we will first briefly
examine how indeterminism enters into the supposedly deterministic
laws of classical physics.
The classical laws of physics are based on algorithms that tell how
the state of a physical system changes with the passage of time. The
state of the system is defined by a collection of numbers called para-
meters. For a simple pendulum, two parameters defining the position
of the pendulum bob and its momentum are sufficient. For many ap-
plications it is sufficient to use such a simple approximation, based on
a lew parameters.
However, an exact Newtonian description would require assigning
parameters to billions and billions of individual molecules. (Thus Avo-
gadro's constant, the number of molecules in a gram-molecular weight
of a substance, is about 6 followed by 23 zeros.) Strictly speaking,
molecules should be described using quantum mechanics, which I will
discuss in Chapter 4. But whether we use quantum theory or classical
Newtonian physics, it is clear that no known computer or process of
computation will enable us to predict what billions and billions of
molecules will do. Physicists have taken it as a matter of faith that all
of the molecules in nature move according to their equations, and in
this sense, physics can be viewed as a branch of theology.
Given the state of a physical system at one moment in time, the
classical laws of physics include formulas for computing how the system
will change over an infinitesimal time interval. By working forward
interval by interval with these formulas (using the methods of calcu-
lus), you can compute how the system changes as time passes. In
principle, if you know the state of the system at one time, you can
compute exactly what the system will do at any future time. This is
called determinism.
 33

Laplace summed this up in 1799 by saying that "if we conceive of

an Intelligence who, for a given moment, embraces all the relations of
beings in this Universe, It would be able to determine for any instant
of the past or future their respective positions, motions, and generally
their effects" (Hahn, 1967, p. 17). Laplace is clearly thinking theologi-
cally, but it can be argued that this mode of thinking actually encour-
aged the development of physics by giving scientists the boldness to
seek a rational understanding of the universe as a whole. The same
mode of thinking, applied to the idea of virtual reality, may also lead to
useful insights into the role of consciousness in nature.
Laplace's hypothetical Intelligence could deduce the future from
the present. However, we now know that the future of many classical
physical systems can be accurately predicted only for a time propor-
tional to the number of decimal places with which the parameters of the
system are known. If you know all of the parameters out to infinitely
many decimal places, then, in principle, you can know the entire future
of the system. But if some of the parameters—or even one—are ac-
curate to only a few decimal places, then your predictions will com-
pletely break down after a short time. This unpredictability has been
celebrated under the label of "deterministic chaos."
The word "chaos" suggests something uncontrollable, but this
turns out to be misleading. A "chaotic" system is susceptible to external
control, but a nonchaotic system cannot be so easily controlled. This
can be illustrated by comparing an old-fashioned pendulum clock with
an automobile on a desert salt flat. If you know roughly the setting of
the clock and how fast it gains or loses time, you can predict its future
behavior easily (relative to some standard clock). But imagine tying
down the gas pedal of a driverless car on a salt flat and letting it go. To
predict the motion of the car you need very accurate information about
how small bumps in the ground are causing its wheels to swing this way
and that.
At the same time, a driver can direct the car according to his will by
making small deflections of the steering wheel. But each time a person
wishes to change the setting of the clock, he has to fully shift the clock's
hands to their new position. Clocklike systems require drastic inter-
vention to change their state, whereas carlike systems can be controlled
by very small interventions. Thus chaos allows for external control,
34

although it simply produces random behavior when left to itself.

The clock was one of the inspiring metaphors behind the mechanis-
tic picture of the universe, but many classical physical systems are more
comparable to the car on the salt flat than to the clock. To see why, it
is useful to study a simple model, called the Baker Transformation.
This is a highly idealized model of a baker kneading a square of dough,
but it illustrates an important phenomenon that occurs in more com-
plex form in many real situations. The sidebar on pages 36 and 37 pre-
sents the details of the Baker Transformation, and the next section
points out their implications. Readers who wish to skip the details can
proceed directly to the next section after reading the first two para-
graphs of the sidebar to find out what the Baker Transformation is.

CONTROL THROUGH AMPLIFICATION

What the Baker Transformation shows is that all information
defining the motions of a point (x,y) in the dough square is stored in the
binary digits defining that point. As time passes, this information is
steadily read out from the x-coordinate of the point and then sent into
oblivion in the higher binary places of the y-coordinate. The digits of
(x,y) are like a script defining all the future motions of the point. In
Figure 4, these scripts were defined for a collection of points so that
the points would spontaneously form a "happy face" at a given time
(square 7). One could similarly write scripts so that these points would
form the letters spelling out Shakespeare's plays, one after another.
The laws of physics of this model are defined by the simple, de-
terministic actions of squashing, cutting, and stacking. But without any
violation of these laws, it is possible for the model to exhibit an in-
telligent message. All that is needed is for the message to be scripted
into high-order binary places of the points that will spell it out.
At this point we need to consider what might be called the hid-
den metaphysics of numbers. In physical science and engineering,
measured quantities, such as the mass of an electron, are normally
written with at most 7 or 8 decimal places. Further decimal digits would
not be considered significant, due to inevitable errors in measurement.
However, the real number system of mathematics assumes that num-
bers have infinitely many decimal places (or binary places, if binary
 35

Figure 4. The Baker Transformation. Suppose the "world" is a square occupied by 100
dots, and suppose that the "laws of nature" are given by the Baker Transformation.
The 12 squares show successive events in this world. The dots seem randomly placed in
square 1, but they move to form a "happy face" in square 7. This pattern was actually
coded into the positions of the dots in square 1, but it became visible after 6 steps as a
result of exponential amplification. (See sidebar on pages 36-37 for more detail.)
36

The Baker Transformation

Consider a baker who is kneading dough. He flattens out the dough, cuts it
into two pieces, and puts one piece on top of the other. Then he repeats this
process. When this is done several times, it has the effect of mixing the
ingredients of the dough (Prigogine, 1980, pp. 220-31).
We can make a simple model of this by using a square to represent a
vertical cross section through the dough. In the flattening step, the square is
squashed into a rectangle that is half as high and twice as wide. Then it is
divided in half by a vertical cut down the middle, and the right half is put on
top of the left half. This produces a square of the same size as the original
square, and the process is repeated.
Let us see what happens to a single point in the dough during this pro-
cess. To do this, it is useful to use binary numbers to locate the point. We all
6
know that the decimal number 3.607 is 3 plus /1o plus Vim plus /1,000. A binary
7

number is similar, but instead of using the 10 digits 0 through 9, only 2 digits
1 0 1 1
0 and 1 are used. Thus 1.1011 is 1 plus /2 plus /4 plus /8 plus /16. In this num-
ber, the dot is called the "binary point," rather than the "decimal point." Also,
the places of the digits could be called binary places rather than decimal
places.
We can let x be the distance of the point from the left side of the square
and y be its height from the bottom of the square. These are just (x,y)
coordinates, with the origin (0,0) fixed at the lower left-hand corner of the
square. Suppose that the side of the original square is one unit in length. We
can write the coordinates x and y in the binary notation. For example, x and
y could be:

x = .101100010111101... y = .110001010010011...

Both x and y are between 0 and 1, and both have infinitely many digits
(which may be 0s or repeating sequences after some point).
During the squashing step, x is doubled and y is divided in half. When a
binary number is multiplied by 2, all the digits shift one place to the left. (This
is similar to what happens when a decimal number is multiplied by 10.)
Likewise, when a binary number is divided by 2, all the digits shift one place
to the right. So the result of the squashing step is:

x = 1.01100010111101... y = .0110001010010011...

The cutting step divides the squashed rectangle into two rectangles.
Since x is now greater than 1, it lies in the right-hand rectangle. In the stacking
step, this rectangle is moved left by 1 and up by .1 so that it sits on top of the

(continued on next page)

 37

(continued from previous page)

1
left-hand rectangle. (Here .1 is /2 in binary notation.) The result of this is that
x and y become:

x = .01100010111101... y = .1110001010010011...

What has happened is that the 1 to the left of the binary point in x has
moved into the slot to the right of the binary point in y. If this digit had been
a 0 instead, then the zero would have been moved. (This case is left to the
reader as a homework problem.)
In effect, the squashing, cutting, and stacking steps simply shift the digits
in x and y. The digits act as though they were on an infinite belt that bends
around the two binary points, and shifts by the space of one digit each time
the squashing, cutting, and stacking operations are performed. The belt can
be illustrated by writing y backwards in front of x and putting two asterisks (*)
in place of the binary points. We can see that the result of all 1 the steps is simply
to shift the ** one digit to the right in the infinite belt of digits:

... 110010010100011 ** 101100010111101...

goes to ...1100100101000111**01100010111101...

Suppose that the kneading operations of squashing, cutting, and stack-

ing are repeated over and over as time passes. Consider a series of digits far
to the right in x, in the binary places numbered from 1,001 to 1,020. These
l m
digits have no measurable effect on x, since they at most contribute 2 to
the magnitude of x. In decimal form, this number is approximated by a deci-
mal point, 300 zeros, and a 1.
After 1,010 operations, the binary points shift over by 1,010 places. The
first 10 of these digits will occupy the first 10 binary places of y, and the second
10 will occupy the first 10 binary places of x. Now the values of x and y are
almost fully determined by these digits. After another 1.010 operations, the
digits will occupy binary places 1,001 to 1,020 in y, and their effect is again
unmeasurably small. Thus the effect of repeating the Baker Transforma-
tion is to bring successive batches of stored information into prominence and
return them to obscurity.

notation is used). The laws of physics are expressed in terms of such

numbers. So does an electron really have a mass specified out to infi-
nitely many decimal places—or for that matter, to billions of decimal
places? Since they are unmeasurable, such vast strings of decimal digits
belong to metaphysics, rather than physics.
38

One might consider replacing a static metaphysical x of infinitely

many digits with a dynamic metaphysical x in which the unobservable
high-order digits are continuously being revised by some agency. In
that case, the motion of the points in the Baker Transformation could
be controlled by that agency. At the same time, this motion would show
no measurable deviation from the physical laws of the model. We can
even say that there is no deviation at all, since a deviation that is
impossible to measure has no real existence in physics.
It is possible to incorporate this hypothetical agency into a man-
made virtual reality that can easily be implemented on a personal com-
puter. The "virtual world" of this VR consists of dots on a square. The
laws of nature in this world are simply the repeated application of the
Baker Transformation, which causes the dots to move about. The
human subject sees the pattern of dots on his screen, and by striking
keys he can change binary digits of the points in the thousandth binary
place or higher. The subject will be able to cause faces or letters to ap-
pear in the square without measurably violating the laws of the system.
I should emphasize that, like all virtual realities, this one has an
interface linking the conscious subject to the virtual world. On the
human side, this interface consists of the screen and keyboard which
enable the subject to interact with the computer. On the side of the
virtual world, it consists of software for updating binary digits of the x's
and y's. The interface is not in the virtual world, and it forms a bridge
between the virtual world and the conscious participant.
Of course, the Baker Transformation is highly idealized and sim-
plified. But similar behavior is shown by many realistic models in phys-
ics. The key feature of the Baker Transformation is that it involves
exponential amplification of small differences with the passage of time.
Consider two points in the square of dough that are very close to each
-1000
other. Suppose that their x coordinates differ by 2 , an unmeasurably
small difference. With each time step, this difference doubles, as long
as the two points are not separated by a cut, and it will reach a mag-
nil mle of 1 after 1,000 steps. This is exponential amplification. (In our
simple VR, "measurement" is limited by the size of the pixels on the
screen, and the limit of 1,000 binary places for updates is overkill.)
In many physical systems, chaotic motion is associated with an
entity called a strange attractor. In physics, an attractor is a set of
 39

points that the state of the system tends to approach as time passes. For
example, the velocity of an object subject to friction will tend to ap-
proach zero as time passes, and thus the point zero is an attractor for
this system.
A strange attractor is typically a very complex set of states. As the
state of the system approaches this set, nearby states will tend to di-
verge from one another exponentially, and they soon exhibit com-
pletely different behavior. Figure 5 gives a simple example showing
how this happens.
Exponential amplification and strange attractors are quite com-
mon in nature. This chapter began with Edward Lorenz's classical

Figure 5. The Henon strange attractor. M. Henon (1976) described a simple transfor-
mation that takes the shaded quadrilateral (top left), stretches it, bends it into a U-
shape, and puts it back into the quadrilateral. When this operation is repeated, the U is
stretched and bent into a double V inside the U, a quadruple U inside the double U, and
so on. In the limit, this becomes an oddly shaped set called a strange attractor. Due to
the stretching, two points that initially are close together in the shaded quadrilateral will
be sent to completely different locations on the strange attractor. This is a simple
example of the kind of exponential amplification that occurs in more complicated
situations in the real world.
40

study of chaos in the atmosphere in 1963. This resulted in the discovery

of a strange attractor in a space of parameters describing atmospheric
motion. Thus we expect meteorological phenomena to be unpredict-
able, and consequently controllable, in principle, by a suitable hidden
agency.
Chaos also shows up in biology. Studies of the behavior of the
human heart suggest that the state of the heart seems to dance chao-
tically about a strange attractor (Goldberger, et al., 1990). As a result,
the beating of the heart continuously varies in an unpredictable man-
ner. Curiously, the heart may begin to beat in a more regular fashion
near the onset of cardiac arrest. It almost seems as though the irregular
beating of the normal heart provides rest for the heart muscle, whereas
highly regular beating is exhausting.
Chaos seems to be a normal feature of the nervous system. Studies
of electroencephalograms of normal individuals conducted at the Free
University of Brussels have shown evidence of chaos. Researchers at
the University of Tubingen in West Germany have similarly found
evidence of chaos in the components of the nervous system controlling
hormone secretion (Goldberger, et al., 1990).
Xu Jing-hua and Li Wei of the Shanghai Institute of Biochemistry
in China have also reported the characteristic signature of chaos in
human E E G s . In addition, these authors analyzed theoretical models
of neuroglia networks which exhibited strange attractors and classical
chaotic behavior. The glial cells in the brain accompany the neurons in
great numbers, but their role in brain functions has tended to remain
obscure. Xu and Li suggest that chaotic neuroglia interactions provide
a source of information. They suggest that this may lead to a "neuroglia
modulation hypothesis of human consciousness and mind formation"
that could account for the creativity of the human brain (Xu and Li,
1986).
Walter Freeman of the University of California at Berkeley has
carried out extensive studies of the olfactory bulb in the brains of
rabbits. He concludes that chaotic dynamics serves an essential role in
the ability of the rabbit to recognize familiar odors and to learn to
recognize new odors. He generalizes his observations by saying that
"Chaos constitutes the basic form of collective neural activity for all
perceptual processes and functions as a controlled source of noise, as
 41

a means to ensure continual access to previously learned sensory

patterns, and as a means for learning new sensory patterns" (Skarda
and Freeman, 1987).

BACK TO THE VIRTUAL BRAIN

Freeman, of course, is not thinking that chaos could be an entry
point of information guiding the action of the brain. But this is certainly
possible, given an agency capable of exerting submicroscopic control.
If we return to the idea of a virtual brain, we can see that deterministic
chaos could be used to inject information coming from outside the
virtual world without any measurable violation of the laws of physics.
In complex, realistic systems this cannot be done as easily as it can be
done for the Baker Transformation. Nonetheless, the control of real
chaotic systems is possible.
For example, researchers at the Naval Surface Warfare Center in
Maryland have shown that chaotic motion in a vibrating magnetoelas-
tic ribbon can be controlled by giving small perturbations to a system
parameter (Ditto, et al., 1990). These perturbations are chosen by
computer calculations in such a way as to induce the ribbon to follow
an unstable orbit embedded in the chaotic attractor.
This was done for two different orbits. By properly choosing the
small perturbations, the system could be induced to follow one orbit
or the other at will. This is clearly significant as a switching technique
for controlling large systems with tiny control signals. Of course, the
signals used to control the ribbon were large enough to be measur-
able. But a virtual reality system could utilize immeasurably small
control signals that would have the same effect. The exponential insta-
bility associated with deterministic chaos can provide the needed
amplification. The full information on system dynamics that is avail-
able to the VR computer can be used to calculate the needed control
signals.
I should finally note that collisions between nearly spherical ob-
jects result in deterministic chaos. Thus molecular motion must every-
where give rise to chaotic effects. The phenomenon of Brownian mo-
tion shows that such effects can be amplified to a macroscopic level,
where they could be used to guide physical systems on a large scale.
42

STATISTICS A N D THERMODYNAMICS

Thus far, we have seen that man-made virtual realities can be

constructed that do not measurably violate the laws of physics. This
suggests that our world could possibly be a virtual reality. However,
there are additional features of modern physics which may seem to rule
out the virtual reality hypothesis. One of these is called the second law
of thermodynamics, and we will now see what it has to say.
Consider a chaotic system that spells out Shakespeare's plays. This
does not contradict the classical laws of physics if the plays are encoded
in the strings of digits defining the state of the system at a given time.
However, such behavior is certainly not expected. One expects random
behavior, but to provide for this it is necessary to add assumptions of
randomness to the laws of physics.
In fact, such assumptions are commonly made to define the direc-
tion of time in physics. The classical equations of motion are time
reversible. This means that if the sequence of events in a physical sys-
tem is run in reverse order, like a home movie played backwards, it still
satisfies the laws of physics. But in nature we see irreversible proc-
esses, such as the scrambling of an egg or the shattering of a glass. Why
do we never see scrambled eggs unscrambling and broken glasses
reassembling and becoming whole again?
The accepted answer is given by a field of physics called thermo-
dynamics or statistical mechanics. Instead of looking at a single state
of a physical system, statistical mechanics looks at collections of many
states called ensembles. Each such collection is associated with a quan-
tity called entropy, which measures the amount of disorder in the col-
lection. The famous second law of thermodynamics states that the
entropy of an isolated physical system tends to increase.
Since the universe is by definition an isolated system, this means
that the amount of disorder in the universe should continuously in-
crease. Indeed, Clausius, one of the 19th-century founders of ther-
modynamics, declared that "The entropy of the universe increases to
a maximum" (Prigogine, 1980, p. 78). This led to the idea of the "heat
death of the universe." The entire universe is supposed to run down
until all energy sources are exhausted and silence reigns over a final
state of featureless thermal equilibrium. The physicist Stephen Hawk-
ing recently described such a future state as follows: "By then all
 43

the stars will have burned out and the protons and neutrons in them
will probably have decayed into light particles and radiation. The
universe would be in a state of almost complete disorder" (Hawking,
1988, p. 151).
But how can we reconcile irreversible increase in disorder with
time-reversible physical laws? To see how this is done, it is convenient
to go back to the Baker Transformation. Like the laws of physics, this
transformation is reversible. To reverse it, simply cut the square in half
with a horizontal cut, put the two rectangles side by side, and squash
them horizontally so they rise up to form a square.
Previously, we studied the motion of a single point under repeated
applications of the Baker Transformation. But now, consider what
happens if we replace the point with the collection of points lying within
a very small circle. Successive transformations stretch this circle hori-
zontally until it becomes a long, thin line. Soon the line crosses the
boundary between the right and left squashed rectangles, and in the
next step it becomes two horizontal lines, one above the other. Then it
becomes 4,8,16, lines and so on. Soon the entire square is covered with
closely spaced fine lines. If we blur these thin lines, spreading them out
to the width of the original circle, then the whole square is completely
covered.
This blurring operation is the key to understanding the second law
of thermodynamics. It is called "coarse graining," and it results in an
irreversible transformation. If we did not blur the fine lines, then
applying the reverse of the Baker Transformation would eventually
bring them back to the original circle. But if we look at the set of
transformed points with blurred vision, we see it spreading out more
and more until it fills the entire square. The blurred sets will not go back
to the original circle under the reversed transformation.
Entropy can be thought of as a measure of how much the blurred
set of points spreads out. When the set is concentrated in the original
small circle, it defines a definite location, and its entropy is low. But as
it spreads out, it loses definition, and the entropy increases. When the
set occupies the whole square it is as indefinite as possible, and the
entropy is at a maximum. So entropy is a measure of lack of definite-
ness.
Now it could be objected that this is all subjective. When we look
44

at a set of points, we may get a blurred image due to the limitations of

our senses and measuring devices. But the actual set of points is
unblurred and is as definite as ever. In fact, this objection is quite true.
The second law of thermodynamics really says that disorder increases
on a measurable level. But on the level of the immeasurably small,
reversibility reigns and the amount of order remains constant.
With this introduction to thermodynamics, let us go back to Shake-
speare's plays. If we use tiny circles instead of points, it may not be
possible for the system to spell out extensive texts. Before they can
form a long series of letters, the tiny circles will have spread out over
the entire square. Whether our vision is blurred or sharp, the system
will tell us nothing about Shakespeare.
The effect of using small circles instead of points, is that we assume
that there is a lack of definiteness in nature. This is related to the idea
of randomness. The circle represents a probability distribution. This
means that there is a definite point, but we don't know exactly where
it is, and thus we say it is one of many points within a circle. Indeed, not
only are we uncertain about where the point is, but nature itself is un-
certain. If we were uncertain, the system might still spell out Shake-
speare's plays. But if nature is uncertain, then this is extremely unlikely
to occur. Suppose the binary digits of x and y are random beyond a
certain point. Then the point (x,y) can fall anywhere in a small square
(and the case of a circle is similar). This makes the appearance of
Shakespeare's plays in the sequences a matter of pure chance.
We see then, that the assumptions of uncertainty made in statistical
mechanics could rule out the kind of hidden agency we have been
considering. If there is an inherent vagueness in nature, then we cannot
postulate a dynamic process which continuously injects information
into nature on the level of the immeasurably small.
But perhaps there is such a dynamic process, and nature does not
suiter from inherent vagueness. In that case, mind-brain interaction
based on the amplification of sub-microscopic information patterns is
a possibility. Such a process might even result in a continual injection
of order into the universe which would prevent its anticipated heat
death. It might also account for the existence of the extensive order
presently existing in the universe. Then again, if randomness and
ordered information are both injected, then we could have an increase
 45

in disorder in some situations and an increase in order in others.

In the brain, the injection of information will be hidden by the large
amounts of information that is already being stored and processed by
neural activity. To show definitely that information is being injected, it
might be necessary to fully understand the information accounting of
the brain and show that the accounts do not balance unless some
additional source of information is assumed. But the extreme complex-
ity of the brain makes this a formidably difficult task. Likewise, to
prove that information is not being injected is also formidably difficult.