Mechanistic and Nonmechanistic Science PDF

MECHANISTIC
AND
NONMECHANISTIC
SCIENCE
An Investigation Into the Nature
Of Consciousness and Form
Richard L.Thompson
BALA BOOKS
Lynbrook, New York
Copyright (5) 1981 by Richard L. Thompson.
All Rights Reserved.
Published by Bala Books, Lynbrook, New York.
Printed in the United States of America.
First Edition.
Library of Congress Cataloguing in Publication Data
Thompson, Richard, L.
Mechanistic and nonmechanistic science.
Bibliography: p.
Includes index.
1. Science—Philosophy. 2. Physics— Philosophy.
3. Consciousness. 4. Form (Philosophy) I. Title.
Q175.P497 501 81-19842
ISBN 0-89647-014-8 (pbk.) A A C R 2
Bala Books, 39 Dawes Avenue, Lynbrook, N . Y . 11563

Dedicated to
His Divine Grace

A . C . Bhaktivedanta Swami Prabhupāda
om ajñāna-timirāndhasya jñānāñjana-śalākayā
caksur unmīlitam yena tasmai śrī-gurave namah
Contents
Introduction 1
P A R T I. CONSCIOUSNESS
Chapter 1. Searching Past the Mechanics of Perception . . . . 13

Chapter 2. Thinking Machines and Psychophysical
Parallelism 27
2.1 How a Computer Works 28
2.2 Artificial Intelligence and Hierarchies
of Function 31
2.3 Subjective Consciousness in Machines
and Humans 33
2.4 Several Nonmechanistic Theories 38
2.5 The Conscious Self as a Complete Sentient
Personality 41
Chapter 3. Dialogue on Consciousness and the Quantum . . . . 49

3.1 A Quantum Mechanical Problem 52
3.2 *What Quantum Mechanics is Really Saying . . . . 55
3.3 *Threading the Labyrinth of Quantum
Epistemology 60
3.4 A Discussion of Contrasting World Views 72
Chapter 4. Karl Popper on the Mind-Body Problem—

A Review 87
PART II. FORM
Chapter 5. Information Theory and the Self-organization

of Matter 97
5.1 *The Theme of Simplicity in the Theories of
Physics 99
5.2 *The Great Complexity of Biological Form . . . . 111
5.3 * Information-theoretic Limitations on the
Evolution of Complex Form 121
5.4 Complex Form and the Frustration of Empiricism . 131
Chapter 6. Chance and the Unity of Nature 143
6.1 Statistical Laws and Their Role in Modern
Physics 144
6.2 The Illusion of Absolute Chance 150
6.3 Chance and Evolution 153
6.4 The Paradox of Unity and Diversity 159
Chapter 7. On Inspiration 169

7.1 The Mechanistic Explanation 172
7.2 Some Striking Examples 174
7.3 The Interaction between Consciousness and
Matter 177
Chapter 8. The Doctrine of Evolution 183

8.1 Evolution: An Invisible Process 184
8.2 The Fossil Record and the Origin of Higher Plants . 186
8.3 The Enigma of Organic Structure 192
8.4 The Resurrection of the Hopeful Monster 197
8.5 Evolution and Negative Theology 202
PART III. CONCLUSION
Chapter 9. The Epistemology of Transcendental

Consciousness 211
9.1 The Process of Bhakti-yoga 214
9.2 Faith, Subjectivity, and Verifiability 217
9.3 The Brain, the M i n d , and the Conscious Self . . . 219
9.4 The Positive and Negative Injunctions of Bhakti . 220
9.5 The Process of Śravanam 222
9.6 The Process of Kīrtanam 224
Appendix 1. * A Discussion of Information Theory 227

Appendix 2. i n f o r m a t i o n Content of the Laws of
Chemistry 233
Bibliography 239
Index 249
Introduction
In an article on the theory of evolution, the biologist John Maynard
Smith declared, "The individual is simply a device constructed by the genes
1
to ensure the production of more genes like themselves." This statement
conveys in a nutshell what modern science has to say about the meaning of
human life. It tells us that the individual person is nothing more than a
machine composed of material elements. This machine has come into
being only because it and the other machines in its ancestral line happened
to be effective at self-duplication in their particular environmental circum-
stances. A l l of this machine's attributes—including its thoughts and feel-
ings, its abilities, and its hopes and desires—are meaningful only insofar
as they contribute to the propagation of the machine's genetic blueprint.
A n d this is meaningless in any ultimate sense, for the genes themselves are
nothing but inanimate molecules.
Smith's statement is by no means exaggerated or atypical. It is a straight-
forward expression of the conclusions of conventional evolutionary theory.
Moreover, if we generalize this statement by allowing it to refer not just
to genes as we know them but to any possible system of physical self-
reproduction, then the statement follows as an unavoidable corollary of the
mechanistic world view of modern science.
The term "mechanistic" refers to the theoretical system of modern phys-
ics, which is based on measurement and calculation. Loosely speaking, the
fundamental premise of physics is that all phenomena are produced by an
underlying stuff called matter. Physicists have developed a number of dif-
ferent theoretical descriptions of matter, and they have still to agree on a
final theory. Yet all of their theories share the following two features:
(1) Matter can be represented by numbers that correspond directly or

indirectly to experimentally measurable properties.
(2) The behavior of matter can be described by mathematical expres-
sions called the "laws of nature."
The world view of modern physics is essentially mathematical, and to a per-

son steeped in this way of seeing things, the mathematical abstractions of
physical theories (such as orbitals, waves, and particles) seem more real
than the tangible phenomena they are used to describe.
Today, research in nearly every scientific field revolves around the
1
1 Introduction
mechanistic premise that all phenomena are due to matter acting in ac-
cord with the laws of nature. In biology this premise implies that living or-
ganisms are combinations of material elements, and that they must have
arisen from earlier states of matter by purely physical processes. Since the
goal of mechanistic theorization is to explain as much as possible by the
natural laws, scientists hypothesize that life developed from matter that
existed originally in a disorganized, nonliving form. This hypothesis has
been systematically elaborated, first in the Darwinian theory of evolution,
and then in the theory of molecular evolution. The first of these theories
deals with the origin of higher species from single-celled organisms, and the
second tries to account for the origin of the first living cells from simple
chemical compounds in a "primordial soup."
In psychology the mechanistic premise implies that mind is merely a
name for certain patterns of electrochemical interaction in the brain. This
means that psychological terms such as "purpose" or "meaning" corre-
spond to nothing more than patterns of behavior that arose as evolutionary
adaptations. The mechanistic premise implies that it is pointless to seek an
absolute sense for such terms or to apply them on a universal scale, for the
universe as a whole consists of nothing but an inexorable flux of physical
actions and reactions. Persons are thus reduced to mere subpatterns of an
inherently meaningless universal pattern.
Although many scientists assert that the mechanistic approach of mod-
ern science is correct, many also admit that it leaves them with a feeling of
dissatisfaction. Thus the physicist Steven Weinberg winds up his account of
the big bang theory by describing human beings as the "more-or-less farci
cal outcome of a chain of accidents reaching back to the first three min-
utes," and he concludes that only the quest for knowledge by physicists like f
himself "lifts human life above the level of farce, and gives it some of the
2
grace of tragedy." The bitterness and disappointment in this conclusion
can also be seen in Bertrand Russell's declaration that, because the mecha-
nistic world view has become so solidly established, "only on the firm foun-
dation of unyielding despair can the soul's habitation henceforth be safely
3
built."
The mechanistic world vision tends to create in sensitive individuals such
a sense of existential despair. It denies the very existence of an absolute
dimension of higher purpose that seems essential for the satisfaction of
the inner self. Of course, some people may argue that if we have no pur-
pose in an absolute sense, we can create our own purpose. Yet this answer,
too, is not satisfying, for if we contemplate such manufactured "purpose"
from the mechanistic viewpoint, we see it dissolve into nothing but a
meaningless juxtaposition of physiochemical events.
Introduction 3
The mechanistic world view also exerts a powerful influence on human

social relationships. The peaceful conduct of day-to-day affairs depends
on people's adherence to certain laws and standards of behavior. These
depend on systems of moral and ethical values that ultimately cannot be
imposed by force, but must be accepted by inner conviction. Yet mecha-
nistic theories can provide no compelling justification for any system of
social values, and they exclude any attempted justification that draws on
nonmechanistic concepts.
A list of higher human qualities might include such items as charity, self-
control, nonviolence, truthfulness, freedom from anger, compassion, free-
dom from greed, modesty, cleanliness, and forgiveness. Such qualities are
certainly conducive to the well-being of human society, but on what basis
can people be induced to value them? Some evolutionists have tried to use
the principle of natural selection, or "survival of the fittest," to explain
higher modes of behavior as convoluted Machiavellian tactics designed
(in an unconscious, mechanistic sense) to further the goal of widely dis-
4
seminating one's genes. Thus far, these arguments have been vague and
unconvincing. Yet even if a convincing case could be made, it seems this
kind of reasoning would not induce people to cultivate higher qualities and
moral principles. Indeed, by creating a sense of cynicism, it might likely
have the opposite effect.
Actually, the mechanistic principles cannot support any line of reasoning
about how people should behave. The mechanistic philosophy implies that
you will simply do whatever your bodily chemistry drives you to do. This
philosophy denies the very idea of the self as a responsible agent with free
will, and thus it also renders meaningless the idea of moral choice.
In the past, people derived their values and their sense of meaning and
purpose in life from traditional, nonmechanistic religious systems. Today
this is still going on, but many people have become convinced that religion
is inferior to science. Mechanistic science has become widely accepted as
the only source of genuine knowledge about the world, and as the domain
of science has been extended, the authority of the traditional religions has
been steadily eroded. Yet religion remains the only source of values and
conceptions of higher purpose. As a result, many people find themselves
in a painful situation: It seems that they can acquire rational knowledge
only at the price of being cast adrift on a trackless sea of fundamental
meaninglessness.
Mechanistic science provides a dynamic method of acquiring more and
more verifiable knowledge about the world, and it has been impressive
because of its prodigious power to explain and manipulate material phe-
nomena. In contrast, Religion seems to depend on the blind acceptance of
4 Introduction
rigid, unverifiable doctrines, many of which have been discredited by mod-

em scientific theories. The world views of the traditional religions are fun-
damentally incompatible with the world view of modern science, and as the
latter vision has gained widespread acceptance, traditional religious con-
ceptions have come to seem more and more antiquated and unrealistic.
For these reasons the credibility of the traditional religions has steadily de-,
clined, and this decline has been accelerated by the corruption, inequity,
and insincerity that have been prominent in many religious establishments.
It has also been aided by the tendency toward compromise, unbridled
speculation, and downright concoction, which has produced a bewildering
welter of conflicting religious sects and philosophies.
With the eclipse of religion and its replacement by the mechanistic phi-
losophy of modern science, human society has been precipitated into a
moral and spiritual crisis. If the mechanistic world view is indeed valid,
then it is hard to see how a satisfactory solution of this crisis can be
achieved. In that case all we can do is try to live with the conscious aware-
ness that life has no intrinsic purpose—or else plunge willfully into
delusion and try to live in a divided house of faith and reason.
Yet there is evidence that the mechanistic world view does not represent
the whole truth. In this book I show on the basis of logic and ordinary evi-
dence that the prevailing theories of physics and biology have serious de-
fects which can be traced to shortcomings in their underlying mechanistic
framework. This negative conclusion suggests that the spiritual crisis of
modern society might be relieved if we could find a truly scientific system of
spiritual knowledge that extends and partly supersedes the current theories
of science. I try to make a positive contribution by introducing such a non-
mechanistic alternative to the mechanistic world view. This system of
knowledge must satisfy the following four criteria:
First, since mechanistic theories have nothing to say about purpose and
personal values, this system must be nonmechanistic. A scientific theory
can be briefly defined as a logically consistent system of statements that can
often be verified by objective observations, and that do not conflict with
observation. A l l of the important theories of modern science can be trans-
lated, at least in principle, entirely into mechanistic terms. In other words,
they can be expressed in the form of statements about measurements and
calculations. Yet there is no reason to suppose that a valid scientific theory
has to be mechanistic. Nonmechanistic scientific theories are possible,
and to establish a sound foundation for spiritual knowledge we must seek
such a theory.
The second criterion is that our system must attribute some dimension
of absolute reality to personality. Unless there is some sense in which per-
Introduction 5
sonality is built into the nature of things, it is not possible to formulate a

satisfactory definition of ultimate purpose. Also, in the absence of some
conception of the individual as an actual being endowed with free will,
there is no meaning to moral and ethical values.
In this connection I should note that there are some nonmechanistic
philosophies—including Buddhism and the monistic philosophies of India
and China—which hold that personality is essentially unreal. Recently
there have been some attempts to reconcile these philosophies with the
5
mechanistic theories of modern physics. However, these efforts do not
achieve the goal that I have in mind. The problem is that in their treatment
of personality, these philosophies are indeed in harmony with the mecha-
nistic world view. They also accept personal existence to be meaningless,
and their aim is to relieve the anxieties of personal life by bringing the
individual to the realization that he fundamentally does not exist.
The third criterion is that the new system must disagree with the existing
theories of modern science to some extent. Some people entertain the hope
of finding a system of spiritual knowledge that harmonizes with the existing
scientific theories. I propose that this hope is unrealistic. The first two
criteria are necessary for an adequate spiritual science, and they cannot be
reconciled with a universal mechanistic system. Yet this does not mean that
the search for a scientifically valid spiritual system is futile. If we closely
examine modern science, we can see that in its pretensions to universality it
has deviated seriously from its observational foundations, and has fallen
into error.
In this book I will discuss two major areas in which any adequate system
of spiritual knowledge must clash with the theories of modern science.
Moth turn out to be areas in which scientists have made unjustifiable
extrapolations of physical theories in an effort to create a universal world
picture. Both are also areas in which a careful examination of theory and
observational evidence shows that the existing scientific picture is seriously
deficient.
Biology is the first of these areas of unjustifiable extrapolation. Modern
biology is founded on the premise that life can be understood completely in
terms of chemistry and physics. No one would deny that many features of
living organisms can be adequately explained by physiochemical models,
and it is reasonable to anticipate that our physiochemical understanding of
biological processes will be greatly extended in the future. Yet no one thus
far has even come close to giving a complete physiochemical analysis of any
living organism.
In modern biology the view that life cannot be fully described by physical
theory is called vitalism. The attitude of many biologists towards vitalism
6 Introduction
is illustrated by the remark, written in a standard biology textbook, that

"those today who may still be prompted to fill gaps in scientific knowledge
6
with vitalism must be prepared to have red faces t o m o r r o w . " It is true that
many vitalistic theories of the past have been faulty, and have been dis-
proven by later scientific findings. Nevertheless, such negative evidence
does not prove that life can be fully explained without recourse to non-
physical principles, and the blustering tone often used by scientists who
make such assertions betrays the weakness of their position. In this book I
will argue that life cannot, in fact, be understood without the introduction
of principles that are not merely nonphysical as we presently understand
this term, but are strictly nonmechanistic.
The field of evolutionary theory is the second area of unjustifiable extra-
polation that I will consider. During the Darwin centennial celebration Sir
Julian Huxley declared, "The evolution of life is no longer a theory; it is a
7
fact and the basis of all our thinking." It is certainly true that a theory of
evolution is an essential ingredient in any universal mechanistic system.
Yet here again, the aggressive tone with which Huxley and others affirm
the present theory of evolution betrays an underlying lack of convincing
proof. In this book 1 will argue that the theory of evolution is not actually
supported by the factual evidence of biology and natural history, and I will
also show that there are fundamental theoretical impediments confronting
any attempt to construct a mechanistic account of the origin of life.
The work of criticizing existing scientific theories is essentially negative,
and its purpose is to clear the way for a positive alternative to the mecha-
nistic world view. But how can we arrive at such an alternative? This
question brings us to the fourth criterion for a scientific system of spiritual
knowledge. If such a system is indeed to provide ultimate standards of per-
sonal meaning, then it must make reference to information that relates to
persons and that is built into the very nature of things. In other words, the
system must have recourse to some universal source of personal direction.
If such a source exists and is accessible to human beings, it stands to
reason that other people may have known about it in the past. Indeed, it
makes little sense to suppose that a genuine source of absolute personal
guidance would remain unknown to human beings throughout history,
only to be revealed at the present time. This consideration greatly simpli-
fies the task of finding a genuine spiritual science. Instead of having to in-
vent such a science from scratch, we should search for it among the many
philosophical and religious systems of the past and present. Our problem
becomes one of recognition rather than one of creation.
A genuine system of spiritual knowledge should have all the characteris-
Introduction 7
tics of a scientific theory. It should provide a logically consistent descrip

tion of reality, and it should entail procedures which can be used to verify
important features of this description. The system should be in agreement
with existing mechanistic theories insofar as they are valid, but it may be
expected to clash with many elements of the modern scientific world view
that rest on unsound speculation and extrapolation. Most importantly, the
system should contain practical methods of obtaining absolute information
about the ultimate meaning and purpose of life.
In this book I will try to make a positive contribution by describing such a
system of spiritual knowledge. This means that I shall introduce a specific
system of theory and practice that has been expressed in a particular lan
guage and handed down in a particular cultural tradition. Since a practical
science must exist in concrete form, it is not possible for me to avoid these
details. Nonetheless, my concern is with general principles that are univer
sally applicable. My purpose is to demonstrate the possibility of a scientific
system of spiritual knowledge by describing an actual example of such a
system. I do not want to pass judgement on other systems or become
embroiled in any kind of sectarian controversy.
I shall describe the system of bhakti-yoga that is expounded in the
Bhagavad-gītā* the Bhāgavata Purāna,9 and other Sanskrit literatures of
India. Bhakti-yoga is the technical name for the philosophy and practical
methodology of a living religious tradition called Vaisnavism. The system
of bhakti-yoga has been taught by many prominent Indian spiritual
teachers, or ācāryas, including Rāmānuja ( A . D . 1017-1137), Madhva
( A . D . 1239-1319), and Caitanya Mahāprabhu ( A . D . 1486-1534). 1 0 1 have
personally learned about bhakti-yoga from His Divine Grace A . C . Bhakti-
vedanta Swami Prabhupāda, and my presentation of this system is based
on his teachings.
The first eight chapters of this book are devoted to a critique of modern
scientific theories and the parallel introduction of basic elements of the
theoretical system of bhakti-yoga. In the ninth chapter I show how these
elements provide the theoretical framework for a practical process of ob
taining absolute personal knowledge. The analysis of current mechanistic
theories is intended to reveal some of their deficiencies, and to show the
need for some kind of nonmechanistic alternative. This analysis does not
prove that the system of bhakti-yoga is the only possible alternative, but it
does show that this system is a reasonable candidate. The validity of bhakti-
yoga can be demonstrated conclusively only by means of the practical ob-,
servational process of bhakti-yoga itself, and this is discussed in the ninth
chapter.
8 Introduction
My discussion of modern scientific theories rests entirely on logic, the

evidence of ordinary experience, and evidence reported in technical jour
nals and other conventional sources of scientific information. Under the
heading of ordinary experience I include the evidence provided by our nor
mal awareness of subjective consciousness. As I will argue in detail later
on, except for our direct experience of consciousness, all of these forms of
evidence can be translated into mathematical language, and thus they are
all mechanistic. Conscious awareness, however, defies representation in
mathematical terms, and thus it is a truly nonmechanistic feature of our
normal experience.
It is not possible to base a nonmechanistic theory entirely on mechanistic
evidence—that is, on patterns of correlation in numerical data. So if non
mechanistic entities and properties are to play more than a vague specula
tive role in a theory, some means must be available for directly observing
them. It is significant, then, that our consciousness, or our inner power of
observation, is of a nonmechanistic character, even though we normally
use it to make observations that can be represented in numerical form.
The nonmechanistic system of bhakti-yoga is, in fact, based on the prin
ciple that the scope of our conscious awareness can be greatly extended.
According to the philosophy of the Bhagavad-gītā, conscious personality is
the irreducible basis of reality. There are two principle types of conscious
being: the one universal Supreme Person, or purusottama, and the innu
merable localized conscious selves, or jīvātmās. Just as electrons interact
with an electric field according to certain natural laws, so the jīvātmās in
teract in a natural way with the purusottama, the all-pervading conscious
being. Bhakti-yoga is concerned primarily with the practical study of this
interaction through direct conscious realization, and thus bhakti-yoga can
be thought of as a kind of physics of higher consciousness.
In physics the interaction between an electron and an electric field can
be studied by certain experimental procedures that take advantage of the
principles of electromagnetic theory. Similarly, the interaction between
the individual self and the Supreme Self can be directly studied by exact
ing procedures that take advantage of the properties of these entities. The
ultimate goal of these procedures is to elevate the consciousness of the
individual jīvātmā so that he can reciprocate with the Absolute Person
on a direct, personal level. Once this is accomplished, the person at
tains incontrovertible knowledge of the nature and meaning of his own
existence.
In the first part of this book, I discuss the nature of individual conscious
ness and introduce the concept of the jīvātmā. There is material on the
Introduction 9
subject of artificial intelligence, the classical theories of the mind-body

relationship, and the mind-body theory of Karl Popper. There is also an
imaginary dialogue in the style of Galileo on consciousness and the role
of the observer in quantum mechanics. I argue that consciousness is objec
tively real, that the contents of consciousness can be correlated only in a
very indirect way with the physiochemical states of the brain, and that con
sciousness cannot be explained in mechanistic terms. Some nonmechanis
tic explanation is required, and I propose that the concept of the jīvātmā
provides a simple explanation that is consistent with known facts.
In the second part I discuss the origin of complex form. In their attempts
to arrive at a universal picture of reality, scientists and philosophers have
always been faced with the problem of how to find unity and harmony in
a world of variegated complex forms. I explore this problem in Chapters 5
and 6, which deal with chance, the laws of physics, and the origin of higher
life forms. Chapter 5 is devoted to an analysis of evolutionary processes
by means of information theory. There I argue that in a mechanistic theory
of the origin of life, unity can be attained only by the sacrifice of com
pleteness, and therefore no satisfactory theory of this kind is possible. I ar
rive at a similar conclusion in Chapter 6 by using nontechnical arguments
about the nature of chance. But I go on to show that a unified theory of the
origin of life is possible if we introduce the concept of the all-pervading
superconscious being, as understood in the philosophy of bhakti-yoga.
('hapters 7 and 8 contain a nontechnical discussion of two types of natu
ral forms—the abstract forms of artistic and mathematical ideas, and the
physical forms of living organisms. Chapter 7 deals with the origin of ideas
and focuses on the phenomenon of inspiration, in which fully developed
ideas appear suddenly and unexpectedly in the mind. I use this phenome
non as the starting point for a discussion of the interaction between the
jivātmā and the Supreme Person.
Chapter 8 deals with the conventional neo-Darwinian theory of evolu
tion, and includes a brief analysis of its historical role as a replacement for
the idea of divine creation. I argue that this theory has never been given a
substantial scientific foundation, and that the idea of creation by an abso
lute intelligent being still provides the most reasonable explanation for the
origin of biological form. This is in accordance with the philosophy of
bhakti-yoga, which holds that all manifestations of form are generated by
the Supreme Person.
The concluding chapter contains a brief outline of the practical observa
tional process of bhakti-yoga. This process takes advantage of the funda
mental principles introduced in the first two parts of this book, and it is the
10 Introduction
only way of giving a tangible demonstration of the validity of these princi

ples. I argue that this process is truly scientific, and that it shares with
modern science such features as a theoretical system, necessary and suffi
cient conditions for the success of experiments, and independent evalua
tion and confirmation of results by a community of experts. The process of
bhakti-yoga is also regulated by a theory of knowledge that strictly rules out
unjustifiable speculation and extrapolation, and in this respect bhakti-yoga
is methodologically superior to modern science. Bhakti-yoga also goes be
yond modern science by providing the individual with practical methods of
developing higher cognitive powers that lie dormant in the conscious self.
I will end this introduction by making a brief observation about the level
of technical difficulty of the material in this book. Since the book deals with
some highly controversial issues, I have felt it necessary to present certain
important arguments on a rigorous technical level. At the same time I have
wanted to make the book accessible to the general reader, and so I have
tried to present as much of the material as possible in an undemanding
style.
As a guide to the reader I have marked the more technical sections with
asterisks in the table of contents. To fully follow these sections, the reader
will need some familiarity with mathematics and physics. Nonetheless, the
arguments they contain are essentially simple, and the general reader may
find it worthwhile to skim these sections without worrying too much about
the technical details.
1. Smith, "The Limitations of Evolutionary Theory," p. 239.

2. Weinberg, The First Three Minutes, p. 144.
3. Russell, "A Free Man's Worship," p. 41.
4. Wilson, On Human Nature.
5. Capra, The Tao of Physics, and Zukav, The Dancing Wu Li Masters.
6. Weisz, Elements of Biology, p. 10.
7. Tax and Callender, eds., Evolution After Darwin, p. 111.
8. A . C . Bhaktivedanta Swami Prabhupāda, Bhagavad-gītā As It Is.
9. A . C . Bhaktivedanta Swami Prabhupāda, Śrīmad-Bhāgavatam.
10. Satsvarupa Dāsa Gosvamī, Readings in Vedic Literature, pp. 50-53.
PART I
CONSCIOUSNESS
Chapter 1
Searching
Past the Mechanics
Of Perception
The life sciences are now dominated by the idea that life can be com-
pletely understood within the framework of chemistry and physics. Those
who subscribe to this viewpoint say that we can explain all features of life—
from the metabolic functioning of cells up to the mental phenomena of
thinking, feeling, and willing—as the consequences of underlying chemi-
cal processes. This viewpoint has become so pervasive that it is generally
presented in biology courses as the only valid understanding of life. Thus,
in textbook after textbook we read that "life means chemical and physical
1
organization," and that "all of the phenomena of life are governed by,
2
and can be explained in terms of, chemical and physical principles."
Yet despite the popularity of this view, we can point to at least one
feature of life—the phenomenon of conscious awareness—that is not
amenable to a molecular explanation. The basic phenomenon of con-
scious awareness is the most immediate aspect of our experience, and it is
automatically presupposed in all our sensations, feelings, and thought
processes. Yet even though consciousness certainly exists and is of cen-
tral importance to our lives, the current theoretical framework of biologi-
cal and physical science cannot even refer to consciousness, much less
explain it.
To see this, let us examine the process of conscious perception through
the eyes of modern science. Our examination will take us through several
levels of successively increasing detail, and at each level we will try to as-
certain whether our scientific picture of reality sheds any light on the nature
of consciousness.
First let us consider a man observing a physical object—in this case, a
thermometer. Figure 1 depicts the operation of the man's sense of sight on
the grossest biological level. The process of perception begins when light
reflected from the thermometer is focused on the retina of the man's eye,
forming an inverted image. This light induces chemical changes in certain
retinal cells, and these cells consequently stimulate adjacent nerve cells to
transmit electrical impulses. These cells in turn stimulate activity in other
13
14 Searching Past the Mechanics of Perception
Figure 1. The process of perception begins when light from an object enters the eye
and is focused on the retina. This light stimulates a series of neurochemical reac-
tions that ultimately reach the brain as a systematic pattern of pulses. These in turn
give rise to an exceedingly complex set of electrochemical reactions within the
brain itself. By studying these processes we may learn a great deal about human
behavior. But will such studies enable us to understand the nature of conscious
perception?
nerve cells, and a systematic pattern of pulses is transmitted down the op-
tic nerve. The image of the thermometer is now encoded in this pattern
of pulses.
When these pulses reach the brain, a very complicated response occurs,
involving many electrochemical actions and reactions. Although scientists
at present do not know the details of this brain activity, they are nonethe-
less in substantial agreement about the basic phenomena involved. When
the impulses streaming down the optic nerve reach the brain, they modify
the overall pattern of chemical concentrations and electrical potentials
maintained by the brain's vast network of nerve cells. This pattern is be-
lieved to represent in coded form the specific content of the man's thoughts
and sensory impressions. As time passes, the physiochemical transforma-
tions of this pattern give rise to sequences of electrical impulses that
emerge from the brain along various motor nerves, and these impulses in
Searching Past the Mechanics of Perception 15
turn evoke corresponding sequences of muscular contractions. These or-

ganized contractions constitute the man's gross external behavior, which
may include spoken reports of his sensations, such as "I am seeing a
thermometer."
At this point in our investigation, we can understand how descriptions of
this kind may, at least in principle, shed light on a person's external be-
havioral responses to environmental stimuli. We can easily imagine con-
structing a machine involving photocells and electronic circuitry that would '
respond to a red light by playing a tape recording of the statement "I am
seeing a red light." On a more sophisticated level, we can visualize a com-
puter that will analyze the images produced by a television camera and gen-
erate spoken statements identifying various objects. Thus although we are
grossly ignorant of the actual physical transformations occuring in the
brain, we can at least conceive of the possibility that these may correspond
to processes of symbol manipulation analogous to those that take place in '
computers. We can therefore imagine that the man's statement, "I am see-
ing a thermometer," is generated by a computational process physically
embodied in the electrochemical activity of the nerve cells in the brain.
But all this tells us nothing about the man's conscious perceptions. Our
description of the image formed on the retina of the man's eye says not hing
about the conscious perception of that image, nor do scientists suppose that
conscious perception takes place at this point. Likewise, the statements
that light-sensitive cells in the retina have been stimulated and that se-
quences of nerve impulses have been induced convey nothing at all about
the actual subjective experience of seeing the thermometer.
Many scientists feel that conscious perception must take place in the
brain. Yet our description of the brain, even if elaborated in the greatest
possible detail, would consist of nothing more than a list of statements
about the electrochemical states of brain cells. Such statements might
have some bearing on patterns of behavior, but they cannot explain
consciousness, because they do not even refer to it.
At this point one may argue that since consciousness is subjective, we
cannot use the word consciousness in scientific statements describing ob-
jective reality. One might point out that while we can observe a man's be-
havior and measure the physical states of his brain, we could not possibly
find any measurable evidence of his so-called consciousness. According to '
this idea, the man's statements about his conscious perceptions are simply
electrochemical phenomena that require physical explanation, but to say
that consciousness exists in any real sense is meaningless.
Each of us can refute this argument by considering the matter in this
Figure 2. Can patterns of neural excitation account for the conscious experience of
seeing?
way: The reality of my own conscious perceptions is certainly undeniable,

and my understanding of all other aspects of reality depends on this basic
fact. Thus I know by direct perception that consciousness exists in me, and
it is also perfectly justifiable to suppose that other beings like me have simi-
lar conscious experiences. There is no need to embrace the futile and ab-
surd viewpoint of solipsism, which holds that I am the only conscious being
and that all others occupy a lesser status as mere automatons. Conscious-
ness, therefore, exists as a feature of objective reality, and any scientific
account of reality that fails to explain it is incomplete.
If consciousness exists but the level of biological description we have
thus far considered does not refer to it, then how can we understand con-
sciousness in terms of our existing scientific world view? The mere asser-
tion that neural impulses "generate" consciousness does not constitute an
explanation, for it offers no conception of any connection between im-
pulses and our conscious perceptions. Our only recourse is to examine the
structures and processes of the brain more closely, with the hope that a
deeper understanding of their nature will reveal such a connection.
Figure 2 presents a closer view of some of the neurons in the brain, and
I'igure 3 depicts some of the minute structures which constitute the inter-
nal machinery of neurons, and of cells in general. When we examine living
cells closely, we find many intricate structures known as organelles. Just
as we can describe the functions of the gross body in terms of the combined
actions of its many component cells, so in principle we can describe the
functions of the cells in terms of these subcellular components. Yet this
does not help us in our attempt to understand consciousness, for it merely
leads to a more complicated account of bodily behavior. As before, there is
no reference to the conscious experience of seeing.
Let us go deeper. What is the essential nature of the cellular organelles?
As we earlier pointed out, the nearly unanimous opinion of modern bio-
chemists is that one can understand all biological structures as combina-
tions of molecules, and all biological processes as the consequences of
Figure 3. A magnified image of a single nerve cell, showing various kinds of

organelles.
molecular interactions. Figure 4 depicts the three-dimensional structure of

a globular protein, one of the many kinds of complex molecules found in
the body. Organic chemistry describes the structure of such molecules in
terms of three-dimensional arrangements of atoms, and molecular inter-
actions in terms of the formation and dissolution of chemical bonds, or
interatomic links.
Figure 4. The three-dimensional structure of myoglobin, a large biological

molecule. Can knowledge of the configurations of such molecules convey anything
about the nature of our conscious awareness? [Redrawn with permission from
R. E. Dickerson, "X-ray Analysis and Protein Structure," p. 634.]
Biochemists have found that living cells contain many different kinds of
extremely complex molecules. For example, the Escherichia coli bacte-
rium, one of the simplest unicellular organisms, is said to contain some two
to three thousand different kinds of proteins, each of which consists of
3
thousands of individual atoms. A complete molecular description of a
single cell would therefore be enormously complex, and, in fact, scientists
have not yet come close to providing such a description, even for the E.
coli bacterium.
Yet however complex it might be, a description on this level would con-
sist of nothing more than a long list of statements about the making and
breaking of chemical bonds. Such a list could give us no greater insight into
the nature of consciousness than any of the higher-order descriptions we

have considered thus far. In fact, lists describing patterns of bonds and lists
describing trains of nerve impulses are equivalent, in the sense that both
say nothing about conscious experience.
Can we find the insight we are seeking by taking a closer look at the
atomic structure of molecules? In Figure 5 we see a diagram representing
Figure 5. An electron density map showing the structure of one of the helical
segments of the myoglobin molecule. [Redrawn with permission from R . E .
Dickerson, "X-ray Analysis and Protein Structure," p. 639.]
the spatial distribution of electrons within an organic molecule. Those who

subscribe to the modern scientific world view claim that we can completely
understand atoms and molecules in terms of the interactions of subatomic
particles such as protons, neutrons, and electrons.
The branch of science that deals with these interactions is known as
quantum mechanics, and it describes subatomic phenomena in terms of
mathematical equations, such as the one depicted in Figure 6(a). Although
diagrams such as Figure 6(b) can depict some features of the solutions to
these equations, their solutions are generally impossible to represent picto-
lially. We might wonder, therefore, whether some deep insight into the
abstract mysteries of these fundamental physical equations might finally
enable us to grasp the nature of consciousness.
Unfortunately, however, this hope must meet with disappointment. If
we study the essential nature of these mathematical equations, we find that
they amount to nothing more than codified rules for the manipulation
Figure 6. A quantum mechanical diagram describing the electronic structure of

a single hydrogen atom is depicted in (b). These structures are deduced from
equations such as the one shown in (a).
of symbols. Such symbols, in turn, are simply marks drawn from a finite
alphabet. They may be represented by the internal states of an electronic
computer or by marks on a piece of paper, but they are essentially arbitrary.
Thus Figure 7 gives us a glimpse of the ultimate appearance of a funda-
mental quantum mechanical description of nature when reduced to its ele-
mental constituent terms. In this figure the alphabet of marks consists of
0,1,2, . . . , 9 , A , B , C F, and the rules for their manipulation are ex-
pressed in terms of the internal language of a particular computer. These
rules simply describe certain ways of rearranging patterns of marks to
create new patterns. Finally, in Figure 8, we reach the end of our investiga-
tion of the scientific world view. Here we find both patterns of marks and
the rules for their manipulation encoded as strings of ones and zeros.
At this point we meet with final frustration in our effort to understand
consciousness in terms of modern scientific conceptions. At each stage of
our investigation we have been confronted with a set of symbols that refer
to repeating patterns in the stream of events we observe with our external
4A00 DD21004B 00130 4A11 DD7E02 00180 4A20 DD7701 00230

4A04 FD21104B 00140 4A14 FD8E01 00190 4A23 DD7E00 00240
4A08 DD7E03 00150 4A17 DD7702 00200 4A26 FD8E00 00250
4A0B FD8603 00160 4A1A DD7E01 00210 4A29 DD7700 00260
4A0E DD7703 00170 4A1D FD8E02 00220 4A2C C32C4A 00270
Figure 7. The reduction of mathematical equations to fundamental rules for symbol

manipulation. Can consciousness be defined in terms of combinations of such
rules?
senses. Thus we began our investigation by describing a man with symbols

like retina and optic nerve, which refer to observable features of gross
anatomy. Now we have ended up with an abstract description in which our
symbols refer to mathematical constructs, or even to elementary rules for
manipulating arbitrary marks on paper. At each successive level of exami-
nation, our symbols failed to refer to consciousness, and, if anything, the
symbols on each successive level seemed more unrelated to the world of
our subjective experience than those on the level above it.
How, then, are we to understand consciousness? Although we know by
direct perception that consciousness exists, we have seen that the method-
ology of modern science does not allow us even to refer to consciousness,
and thus this methodology cannot reveal anything about it. Of course,
some measurable physiological phenomena may be correlated more or
less closely with the contents of a person's consciousness, but a detailed
account of these phenomena does not have anything to say about con-
10110110101001100000111010101111110010011
10100011100111000101100100010000111101001
00011011010100010110101111001010100011111
10010000100101010101100011110100011101010
00000000010011010010101001010011111010110
Figure 8. The ultimate reduction of scientific description to patterns of marks—
in this case ones and zeros—that correspond directly or indirectly to the outcomes
of physical measurements. Although consciousness is the primary feature of our
existence, by its very nature it eludes description in these terms.
sciousness itself. When contemplating records of measurements made on

another person, we have no real need to bring in the concept of conscious-
ness at all, and indeed, the entire science of behavioral psychology is dedi-
cated to the principle that people can be fully described without reference
to consciousness. Our only basis for inferring the presence of consciousness
from quantifiable features of another person is that we find such features
associated with consciousness in ourselves, and by analogy we assume that
this is true of others also.
Now, we have used the term "consciousness" to refer to the reality of our
personal experience of sensations, thoughts, and feelings. We have em-
phasized that while these experiences are certainly subjective, they are
nonetheless factual and are therefore part of the total reality of the world.
We should note that while this fact has gone essentially unrecognized in
the prevailing world view of modern science, many scientists and philoso-
phers have pointed it out from time to time. For example, the physicist
Eugene Wigner has argued that "there are two kinds of reality or existence:
the existence of my consciousness and the reality or existence of every-
4
thing else. This latter reality is not absolute but only relative." Wigner is
observing that external, measurable phenomena are known to him only by
virtue of his consciousness, and thus consciousness is, if anything, more
real than these phenomena.
Another well-known scientist who has recognized the reality of con-
scious experience is Thomas Huxley. After noting that the materialistic
philosophers of his day attributed all phenomena to matter and force,
Huxley wrote that "it seems to me pretty plain that there is a third thing in
the universe, to wit, consciousness, which . . . I cannot see to be matter or
5
force, or any conceivable modification of either." Huxley held that con-
sciousness must be an objective feature of universal reality that is somehow
associated with the bodies of individual persons, but that does not fall
within the categories of matter and energy. He proposed that conscious-
ness is an "epiphenomenon"—a wholly nonmaterial phenomenon that is
somehow generated by certain material interactions, but that does not
affect them in turn.
According to Huxley's idea, consciousness is real but can only be per-
ceived privately. Each person's consciousness is a causal dead end that is
influenced by material processes in the brain, but cannot influence any
conceivable material measuring apparatus. Furthermore, although con-
sciousness is self-referential and must somehow act on itself to perceive
itself, there is no provision in Huxley's scheme for the consciousness of
one individual to directly influence that of another. If Huxley is correct, we
cannot possibly learn more about consciousness than we already know.

The consciousness of each person must remain the sole perceiver of itself,
and the passive observer of the physiochemical processes of the b o d y —
processes over which it has no influence, even when the body is writing
essays on the nonphysical nature of consciousness.
We would like to suggest that while Huxley was correct in pointing out
that consciousness is something distinct from matter, his conception of
consciousness as an epiphenomenon is too restrictive. In this chapter we
would like to outline an alternative model in which consciousness is granted
attributes enabling us to directly study it as a nonphysical, honquantifi-
able feature of reality. Our sources of information for this model are the
Bhagavad-gītā and other literatures from the Vedic tradition of India. 6
From these sources we will draw a few basic points about the nature of con
sciousness that can be thought of as axioms for our model. Here our objec-
tive will not be to prove the truth of the model, but rather, to gain some
idea of what may possibly be true. As we shall see, however, our axioms
will provide hints as to how the model can be investigated. As is true with
any scientific theory, it is only by the positive results of such investigation
that the value of the model can be established.
In our model we postulate that the consciousness of each person is due to
the presence of a distinct nonphysical entity called the self, or jīvātmā.
Some of the attributes of the jīvātmā are summarized in the following list:
1. The jīvātmā is the elementary unit of consciousness.

2. There are innumerable jīvātmās, and they can be neither created
nor destroyed.
3. A l l jīvātmās are qualitatively equal.
4. The jīvātmās are not numerically describable, although they can ex-
hibit quantifiable attributes such as position in space.
5. The jīvātmās obey higher-order psychological laws involving quali- ^
ties and modes of activity that are not amenable to mathematical
formulation.
6. The jīvātmās do not interact with matter according to the known laws
of physics, such as the law of gravity or the laws of electromagnetism.
7. The jīvātmās possess self-reflective conscious awareness, and they
possess innate senses capable of perceiving both matter and other
conscious entities.
8. The jīvātmās tend to be associated with material bodies, but they are
not dependent on matter and are fully capable of functioning without
material connections.
The first of these axioms introduces the self as an irreducible, natural

entity analogous to the idealized fundamental particles that physicists have
sought as the basis of matter. The essential idea here is that consciousness
is an inherent feature of reality and not merely a name for certain physio-
chemical processes. Axiom 2 indicates the individualized nature of con
sciousness, although this axiom is, of course, too strong to be proven by
the mere existence of such individuality. Although similarly difficult to
prove, Axiom 3 provides the basis for a philosophy of universal equality by
declaring the innermost selves of all sentient beings to be equal.
Axioms 4 and 5 express the basic conclusion we deduced from our anal
ysis of the process of perception. Consciousness involves qualities and at
tributes, such as the color red, that cannot be numerically described, and
thus the jīvātmā, or elementary unit of consciousness, also cannot be nu
merically described. Since the days of Newton, many scientists have been
greatly inspired by the idea that all significant features of reality can be rep
resented by patterns of numbers. Yet this idea is simply a convenient work
ing hypothesis and not a necessary truth. Here we would like to introduce
the idea that the world may contain many qualities, features, and entities
that cannot be discussed quantitatively. We should stress, however, that a
numerically indescribable entity may have some quantifiable features, as
indicated in Axiom 4. By "numerically indescribable" we only mean that
all attempted numerical descriptions must be significantly incomplete. We
do not mean that they cannot apply at all.
Axioms 6, 7, and 8 give us a hint as to how our model of the self can be
practically investigated. The essential ingredient for such an investiga
tion is provided by Axiom 7, which states that the jīvātmā is equipped
with its own innate senses. In our analysis of the process of perception, we
studied the material sense of vision, and saw how information depicting
an observed object eventually became represented in the brain in the form
of nerve impulses. We were never able, however, to explain how such
patterns of pulses give rise to the conscious perception of the object.
In our model this is explained by postulating that the jīvātmā possesses
its own senses, but is using them only to pick up information provided by
the sensory machinery of the body. The situation of the embodied jīvātmā
can be compared with that of a person flying an airplane through dense
clouds on instruments. In that situation the person can obtain only a very
limited picture of his surroundings from devices such as the radar screen
and the altimeter, even though he still possesses his normal senses and, in
fact, is using them to observe these instruments.
Since the sensory apparatus of the body is composed of matter, this ap
paratus can provide information only about configurations of material .
energy and their transformations. This information may be used to make
indirect inferences about the jīvātmā itself, but cannot directly reveal any
thing about it. Yet Axiom 7 opens up the possibility that the jīvātmā may be
able to directly obtain information about other jīvātmās by using the full *
power of its own natural senses. If this could be done, the perception of
consciousness would not be a strictly subjective affair. A group of persons
who could perceive one another with awakened senses as conscious be
ings, could discuss and study consciousness with the same objectivity that
scientists now bring to bear on the study of inanimate matter.
In this chapter our purpose is only to indicate the possibility of such a di
rect study of consciousness. Practical methods of carrying out such a study
will be discussed in Chapter 9. Here we will close by noting an interesting
feature of one of these methods—the method of bhakti-yoga, which is de
scribed in Sanskrit literatures such as the Bhakti-rasāmrta-sindhu of Śrlla
Rūpa Gosvāml. 7 This method involves the chanting of mantras such as
Hare Krsna, Hare Krsna, Krsna Krsna, Hare Hare

Hare Rāma, Hare Rāma, Rāma Rāma, Hare Hare.
According to the system of bhakti-yoga, such chanting can purify the

jīvātmā of the effects of involvement with matter, and thereby awaken the
jīvātmā's natural sensory capacities. The interesting feature of this chant
ing is that the practitioners of bhakti-yoga understand the words Hare,
Krsna, and Rāma to be symbols with an absolute meaning lying outside the
realm of numerical describability.
In the quantitative science of physics, we encounter numbers that are
2
ostensibly absolute. For example, the fine structure constant, ћcle =137,
is believed by some physicists to represent an absolute, unchangeable
feature of reality. 8 Now, we can reason that in a direct approach to the
study of consciousness, it would be necessary to employ concepts, and even
symbols, which are more than mere patterns of marks, but which possess
a numerically indescribable dimension. To be fully appreciated, such
symbols would have to be directly perceived by the senses of the jīvātmā,
and not merely processed through the sensory apparatus of the body. It
is therefore interesting to note that the system of bhakti-yoga makes use
of absolute symbols of this kind—symbols that could be referred to as
transcendental constants of nature.
Notes
1. Elliot and Ray, Biology, p. 67.
2. Villee and Dethier, Biological Principles and Processes, p. 12.
3. Watson, Molecular Biology of the Gene, p. 69.
4. Wigner, "Two Kinds of Reality," p. 251.
5. Huxley, Essays on Some Controverted Questions, p. 220.
6. A . C . Bhaktivedanta Swami Prabhupāda, Bhagavad-gītā As It Is.
7. A . C . Bhaktivedanta Swami Prabhupāda, The Nectar of Devotion.
8. Dirac, "The Evolution of the Physicist's Picture of Nature," pp. 48-49.
Chapter 2
Thinking Machines
And
Psychophysical
Parallelism
Science fiction writers often try to solve the problems of old age and
death by taking advantage of the idea that a human being is essentially a
complex machine. In a typical scene, doctors and technicians scan the head
of the dying Samuel Jones with a "cerebroscope," a highly sensitive instru-
ment that records in full detail the synaptic connections of the neurons in
his brain. A computer then systematically transforms this information into
a computer program that faithfully simulates that brain's particular pattern
of internal activity.
When this program is run on a suitable computer, the actual personality
of M r . Jones seems to come to life through the medium of the machine.
"I've escaped death!" the computer exults through its electronic phoneme
generator. Scanning about the room with stereoscopically mounted TV
cameras, the computerized " M r . Jones" appears somewhat disoriented in
his new embodiment. But when interviewed by old friends, " h e " displays
Mr. Jones's personal traits in complete detail. In the story, M r . Jones lives
again in the form of the computer. Now his only problem is figuring out
how to avoid being erased from the computer's memory.
Although this story may seem fantastic, some of the most influential
thinkers in the world of modern science take very seriously the basic prin-
ciples behind it. In fact, researchers in the life sciences now almost uni-
versally assume that a living being is nothing more than a highly complex
machine built from molecular components. In the fields of philosophy
and psychology, this assumption leads to the inevitable conclusion that
the mind involves nothing more than the biophysical functioning of the
brain. According to this viewpoint, we can define in entirely mechanistic
terms the words we normally apply to human personality—words like
consciousness, perception, meaning, purpose, and intelligence.
Along with this line of thinking have always gone idle speculations about
the construction of machines that can exhibit these traits of personality.
27
28 Thinking Machines and Psychophysical Parallelism
But now things have gone beyond mere speculation. The advent of modern
electronic computers has given us a new field of scientific investigation
dedicated to actually building such machines. This is the field of artificial
intelligence research, or "cognitive engineering," in which scientists pro-
ceed on the assumption that digital computers of sufficient speed and
complexity can in fact produce all aspects of conscious personality. Thus
we learn in the 1979 M . I . T . college catalogue that cognitive engineering
involves an approach to the subjects of mind and intelligence which is
"quite different from that of philosophers and psychologists, in that the
1
cognitive engineer tries to produce intelligence."
In this chapter we shall examine the question of whether it is possible for
a machine to possess a conscious self that perceives itself as seer and doer.
Our thesis will be that while computers may in principle generate complex
sequences of behavior comparable to those produced by human beings,
computers cannot possess conscious awareness without the intervention of
principles of nature higher than those known to modern science. Ironically,
we can base strong arguments in support of this thesis on some of the very
concepts that form the foundation of artificial intelligence research. As
far as computers are concerned, the most reasonable inference we can
» draw from these arguments is that computers cannot be conscious. When
applied to the machine of the human brain, these arguments support a
nonmechanistic understanding of the conscious self.
2.1 How a Computer Works

We shall proceed by raising some questions about a hypothetical com-
puter that possesses intelligence and conscious self-awareness on a human
level. This computer need not duplicate the mind of a particular human
being such as the M r . Jones of our story, although this is also an interesting
possibility to consider. We will simply assume that it experiences an aware-
ness of thoughts, feelings, and sensory perceptions that is comparable to
our own.
First, let us briefly examine the internal organization of our sentient
computer. Since it belongs to the species of digital computers, it consists
of an information storehouse, or "memory," an apparatus called the cen-
tral processing unit, or C P U , and various devices for the exchange of
information with the environment.
The memory is simply a passive medium used for recording large amounts
of information in the form of numbers. A typical computer memory can
be visualized as a series of labeled boxes, each of which can store a num-
ber. Some of these boxes normally contain numerically coded instructions
How a Computer Works 29
that specify the program of activity of the computer. Others contain data
of various kinds, and still others are used to store the intermediate steps
of calculations. These numbers can be represented physically in the
memory as patterns of charges on microminiature capacitors, patterns of
magnetization on arrays of small magnets, and in many other ways.
The central processing unit performs all of the computer's active op-
erations. It is capable of performing a fixed number of simple operations
of symbol manipulation. These operations typically include the follow-
ing steps: First, a coded instruction identifying the operation to be per-
formed is obtained from a specified location or "address" in the memory.
According to this instruction, other data may also be obtained from the
memory. Then the operation itself is performed. This may involve reading
a number into the memory from an external device (input), or transmitting
a number from the memory to such a device (output). It may involve the
transformation of a number according to some simple rule, or the shifting
of a number from one memory location to another. Finally, the operation
will always involve the selection of a memory address where the next coded
instruction is to be sought.
The activity of the computer consists of nothing more than the repetition
of steps of this kind, one after another. The specific operations to be exe-
cuted are specified by the instruction codes stored in the passive memory
record. The function of the CPU is simply to carry them out sequentially.
Like the memory, the CPU can be constructed physically out of many dif-
ferent kinds of components, ranging from microminiature semiconductor
junctions to electromechanical relays. The functioning of the CPU is deter-
mined only by the logical arrangement of these components, and not by
their particular physical constitution.
The operation of a computer can be understood most easily by consider-
ing a simple example. Figure 1 illustrates a program of computer instruc-
2
tions for calculating the square root of a number. The thirteen numbered
statements correspond to the list of coded instructions stored in the com-
puter's memory, but here they are written out in English for clarity. There
are also five boxes labeled (1) through (5) that correspond to areas in the
memory intended for the storage of data and intermediate computational
steps. To simulate the operation of the computer, begin by placing a num-
ber, such as 9, in box (1). Then simply follow the instructions one at a time.
When you have completed the last instruction, the square root of your
original number will be contained in box (2). In an actual computer, each of
these instructions would be carried out by the C P U . They illustrate the
kind of elementary operations used by present-day computers (although
Figure 1. Computer program for computing the square root of a number. To simu-
late the operation of the computer, place the number in box (1) and follow the
instructions, starting with step 1. When step 13 is completed, the square root of the
number (rounded down to an integer) will be in box (2). (In these instructions,
"increment a number" means to add 1 to it, and "decrement a number" means to
subtract 1 from it.)
they do not correspond exactly to those of any particular computer).

It may seem that the method of finding a square root given in this ex-
ample is both cumbersome and obscure, but it is typical of how computers
operate. The practical applicability of computers is in fact based on the ob-
servation that every fixed scheme of computation that has ever been for-
mulated can be reduced to a list of simple operations of the kind used in the
example. This observation was made by a number of mathematicians in the
1
1930s and 1940s, and is commonly referred to as Church's thesis. It implies
that, in principle, any scheme of symbol manipulation that can be precisely
defined can be carried out by a digital computer of the modern type.
Artificial Intelligence and Hierarchies of Function 31
At this point, let us consider our hypothetical sentient computer. Ac-

cording to the exponents of artificial intelligence, the intricate behavior
characteristic of a human being can be completely described as a highly
complex scheme of symbol manipulation. By Church's thesis this scheme
can be broken down into a program of instructions comparable with our
example in Figure 1. The only difference is that this program will be ex-
ceedingly long and complex, and may run to millions of steps. Of course,
up until now no one has even come close to actually producing a formal
symbolic description of human behavior. However, for the sake of argu-
ment, let us suppose that such a description could be written and expressed
as a computer program.
Let us suppose that a highly complex program of this kind is being exe-
cuted by a computer, and let us see what we can understand about the com-
puter's possible states of consciousness. When such a program is being
executed, the computer's CPU will be carrying out only one instruction at
any given time, and the millions of instructions comprising the rest of the
program will exist only as an inactive record in the computer's "memory."
Now, intuitively it seems doubtful that a mere inactive record could have
anything to do with consciousness. Where, then, does the consciousness of
the computer reside? The CPU at any given moment is simply performing
some elementary operation, such as, "Copy the number in box (1687002)
into box (9994563)." In what way can this be correlated with the conscious
perception of thoughts and feelings?
2.2 Artificial Intelligence and Hierarchies of Function

The researchers of artificial intelligence have a definite answer to this
question that is based on the idea of levels of organization in a computer
program. To illustrate what is meant by levels of organization, we shall
once again consider the simple computer program of Figure 1. Then
we shall apply this concept to the program of our sentient computer and
s e e what light it can shed on the relation between consciousness and the
computer's internal physical states.
Although the square root program of Figure 1 may appear to be a form-
less list of instructions, it actually possesses a definite structure that is out-
lined in Figure 2. This structure consists of four levels of organization. On
I he highest level, the function of the program is described in a single sen-
tence that uses the symbol "square root." On the next level, the meaning of
this symbol is defined by a description of the method used to find square
roots in the program. This description makes use of the symbol "squared,"
which is similarly defined on the next level down in terms of another sym-
bol "sum." Finally, the symbol "sum" is defined on the lowest level in terms
of the combination of elementary operations that is actually used to com-

pute sums in the program. Although we have used English sentences in Fig-
ure 2 for the sake of clarity, the description on each level should use only
symbols for elementary operations, or higher-order symbols defined on the
next level down.
These graded symbolic descriptions actually define the program in the
sense that if we begin with level 1, and expand each higher-order symbol we
encounter in terms of its definition on a lower level, we will wind up writing
the list of elementary operations in Figure 1. The descriptions are useful in
that they provide an intelligible account of what is happening in the pro-
gram. Thus, on one level we can say that numbers are being squared, on
another level that they are being added, and on yet another that they are
being incremented and decremented. However, the levels of organization
of the program are only abstract properties of the list of operations given in
Figure 1. When the program is being executed by a computer, these levels
do not exist in any real sense, and only the elementary operations in the list
are actually being carried out by the computer.
In fact, we can even go further and point out that this last statement is not
strictly true. Actually, what we call the elementary operations are them-
selves symbols, such as Increment (3), that refer to abstract properties of
the underlying machinery of the computer. When the computer is oper-
ating, all that is really happening within it are certain transformations of
matter and energy which follow a pattern determined by the computer's
physical structure.
In general, any computer program that performs some complex task can
be resolved into a similar hierarchy of levels of description. Researchers in
1. Find the square root of X.

2. The square root of X is one less than the first
number Y with Y squared greater than X.
3. Y squared is the sum of Y copies of Y.
4. The sum of Y and another number is the result
of incrementing that number Y times.
Figure 2. Levels of organization of the program in Figure 1. The program in Figure

1 can be analyzed in terms of a hierarchy of abstract levels. The level of elementary
operations is at the bottom, and each higher level makes use of symbols (such as
squared) that are defined on the level beneath it.
Subjective Consciousness in Machines and Humans 33
the field of artificial intelligence generally visualize their projected "intel-

ligent" or "sentient" programs in terms of a hierarchy similar to the follow-
ing: On the bottom level the program is described in terms of elementary
operations. Then come several successive levels in which mathematical
procedures of greater and greater intricacy and sophistication are defined.
After this comes a level in which symbols are defined that refer to basic
constituents of thoughts, feelings, and sensory perceptions. This is fol-
lowed by a series of levels involving more and more sophisticated mental
4
features, culminating in the level of the ego, or self.
Here, then, is how the relation between computer operations and con-
sciousness is understood by these researchers: Consciousness is associated
with the higher levels of operation of a "sentient" program—levels on
which symbolic transformations take place that directly correspond to
higher sensory processes and the transformations of thoughts. In contrast,
the lower levels are not associated with consciousness. Their structure can
be changed without affecting the consciousness of the computer as long as
the higher level symbols are still given equivalent definitions. Referring
again to our square root program, this corresponds to the observation that
the process of finding a square root given on level 2 in Figure 2 will remain
essentially the same even if the operation of squaring is defined on level 3
in some different but equivalent way.
If we were to adopt a strictly behavioristic criterion for our use of the
word "consciousness," then this understanding of computerized conscious-
ness might be satisfactory—granting, of course, that a program with the
required higher-order organization could indeed be created. By such a cri-
terion, certain patterns of behavior are designated as conscious and others
are not. Generally, a sequence of behavioral events will have to be quite
long in order to merit the designation "conscious." For example, a long
speech may exhibit certain complex features that identify it as "conscious,"
but none of the words or short phrases that make it up could be long enough
to display such features. Using such a criterion, one might want to desig-
nate a certain sequence of computer operations as "conscious" because it
possesses certain abstract higher-order properties. Then the overall be-
havior of the computer might be analyzed as "conscious" in terms of these
properties, whereas any single elementary operation would be too short
to qualify.
2.3 Subjective Consciousness in Machines and Humans

Yet we are interested not in categorizing conscious behavior, but rather
in understanding the actual subjective experience of conscious awareness.
Figure 3. The relation between

consciousness and the physical
structures of a hypothetical sentient
computer. If we assume that a
computer has subjective experience
comparable to our own, then the
computer's awareness of thoughts
and feelings must be real. Yet in the
computer as we know it, these
thoughts and feelings can only
correspond to higher-order abstract
properties of the computer's
program. Here these properties are
represented within the tinted
section by a hierarchy of symbolic
descriptions. Since these properties
exist only in an abstract sense, and
are not actually present in the
computer's hardware, how can they
correspond to concretely real
subjective experiences?
To clearly distinguish this conception of consciousness from the behavioral

one, we will pause here to briefly describe it and establish its status as a sub-
ject of serious inquiry. By "consciousness" we are referring to the aware-
ness of thoughts and sensations that we directly perceive and know that we
perceive. Since other persons are similar to ourselves, it is natural to sup-
pose that they are similarly conscious. If this is accepted, then it follows
that consciousness is an objectively existing feature of reality that tends to
be associated with certain material structures, such as the bodies of living
human beings.
Now, when a common person hears that a computer can be "conscious,"
his natural tendency is to interpret this word in the sense that we have just
described. Thus, he will imagine that a computer can have subjective,
conscious experience similar to his own. Certainly this is the idea behind
the fictional stories such as the one with which we began this chapter.
One imagines that the computerized " M r . Jones" is actually feeling aston-
ishment at his strange transformation as he looks about the room through
the computer's TV cameras. This implies that his feeling of astonishment
really exists.
If this is possible for the computerized Mr. Jones, then we are faced with
the situation depicted in Figure 3. On the one hand, the conscious expe-
rience of the computer exists—its subjective experience of colors, sounds,
thoughts, and feelings is an actual reality. On the other hand, the physical
structures of the computer exist. However, we cannot directly correlate
consciousness with the concrete physical processes of the computer, nor
can we relate it to the execution of individual elementary operations, such
as those of Figure 1. According to the artificial intelligence researchers,
consciousness should correspond to higher-order abstract properties of
the computer's physical states—properties described by symbols such
as "thought" and "feeling" that stand at the top of an extensive pyramid
of abstract definitions. Indeed, these abstract properties are the only
conceivable features of our sentient computer that could have any direct
correlation with the contents of consciousness.
Since consciousness is real, however, and these abstract properties are
not, we can only conclude that something must exist in nature that can
somehow "read" these properties from the physical states of the computer.
This entity is represented in Figure 3 by the arrow connecting the real con-
tents of consciousness with higher levels in the hierarchy of abstract sym-
bolic descriptions of the "sentient computer." It must have the following
characteristics:
(1) This entity must possess sufficient powers of discrimination to

recognize certain highly abstract patterns of organization in
arrangements of matter.
(2) It must be capable of establishing a link between consciousness
and such arrangements of matter. In particular, it must modify the
contents of conscious experience in accord with the changes these
abstract properties undergo as time passes and the arrangements

of matter are transformed.
Clearly there is no place for an "entity" of this kind in our current picture of
what is going on in a computer. Indeed, we can only conclude that " i t " must
correspond to some feature of nature that is completely unknown to
modern science.
This, then, is the conclusion that we are forced to adopt if we assume that
a computer can be conscious. Of course, we can easily avoid this conclusion
by supposing that no computer will ever be conscious, and this may indeed
be the case. However, what can we say about the relation between con-
sciousness and the physical body in a human being? On the one hand we
know that human beings possess consciousness, and on the other we are
taught by modern science that the human body is an extremely complex
machine constructed from molecular components. Can we arrive at an un-
derstanding of human consciousness that does not require the introduction
of an entity of the kind described by statements (1) and (2)?
Ironically, if we try to base our understanding on modern scientific
theory, then the answer is no. The reason is that all modern scientific
attempts to understand human consciousness depend, directly or indi-
rectly, on an analogy between the human brain and a computer. In fact, the
scientific model for human consciousness is machine consciousness!
At the present time, most scientists regard the brain as the seat of con-
sciousness. The brain is understood to consist of many different kinds of
cells, each of which is regarded as a molecular machine. Of these, the nerve
cells are known to exhibit electrochemical activities that are roughly
analogous to those of the logical switching elements used in computer cir-
cuitry. Even though at present the operation of the brain is understood
only in very vague and general terms, scientists generally conjecture that
these neurons are organized into an information-processing network that is
equivalent to a computer's.
This leads naturally to the picture of the brain illustrated in Figure 4. In
this picture thoughts, sensations, and feelings must correspond to higher
levels of brain activity that are comparable to the higher organizational
levels of a complex computer program. Just as the higher levels of such a
program are abstract, these higher levels of brain activity must also be
abstract. They can have no actual existence, for all that is actually happen-
ing in the brain are certain physical processes such as the pumping of
sodium ions through neural cell walls. If we try to account for the existence
of human consciousness in the context of this picture of the brain, we can
only conclude (by the same reasoning as before) that some entity described
Figure 4, The relation between

consciousness and the physical
structures of the brain. A person's
subjective experiences and the
physical phenomena of his brain are
both real, but these experiences
cannot be correlated with physical
brain phenomena in a direct, one-
to-one fashion. Rather, the
experiences of thoughts and
feelings must correspond to higher-
order abstract properties of brain
states. (As in Figure 3, these
properties are represented by the
hierarchy of symbolic descriptions
within the tinted section.) Since
these abstract properties have no
concrete existence, and are related
only in a very indirect way with the
physical structures of the brain, how
can they correspond to the person's
actual conscious experiences?
by statements ( 1 ) and ( 2 ) must exist to account for the connection between

consciousness and abstract properties of brain states.
Furthermore, if we closely examine the current scientific world view,
we can see that the conception of the brain as a computer does not simply
depend on some superficial details of our understanding of the brain. On a

deeper level we can see that this conception follows necessarily from the
fact that the scientific world view is mechanistic. Mechanistic explanations
of phenomena are, by definition, based on systems of calculation. By
Church's thesis, all systems of calculation can, in principle, be represented
in terms of computer operations. In effect, all explanations of phenomena
in the current scientific world view can be expressed in terms of either
computer operations or some equivalent symbolic scheme.
This implies that all attempts to describe consciousness within the basic
framework of modern science must lead to the same problems that we have
5
encountered in our analysis of machine consciousness. Some entity of the
kind described in statements (1) and (2) will be required to account for con-
sciousness. Yet in the present theoretical system of science, there is nothing
to be found, either in the brain or in a digital computer, that corresponds to
such an entity. Indeed, there could not be, for any mechanistic addition to
the current picture of, say, the brain would simply constitute another part
of that mechanistic system, and the need for entities satisfying (1) and (2)
would still arise. Clearly a nonmechanistic approach to consciousness is
needed.
2.4 Several Nonmechanistic Theories

Over the years, philosophers have developed a number of nonmecha-
nistic theories to account for our conscious experience of thoughts and
feelings. Some of these theories are not compatible with the mechanistic
assumptions of modern science, and others have been specifically designed
to supplement these assumptions without contradicting them. In this
section we will briefly examine some of these theories.
First, let us place the discussion in a formal philosophical perspective,
by noting some of the technical philosophical terms associated with the
strictly mechanistic approach to understanding sentient beings. This ap-
proach is based, in essence, on the idea that a sentient being can be ade-
quately characterized by a mathematical description of its physical states,
or of some aspects of its physical states. Although we have used the broad
terms "behavioristic" and "mechanistic" to designate this approach,
philosophers have actually subdivided it into a number of distinct ver-
sions, with names such as radical behaviorism, logical behaviorism, and
functionalism.
Of these, functionalism stresses the idea that each sentient being can be
characterized by a suitable computer program. Functionalism provides the
philosophical framework for research in artificial intelligence, and it can be
Several Nonmechanistic Theories 39
6
regarded as the most sophisticated of the various forms of behaviorism.
As such, functionalism may seem satisfactory as long as we ignore the exis-
tence of consciousness. But since computer coding says nothing about con-
scious experience, functionalism is completely unable to account for this
essential feature of sentient beings. Thus one advocate of functionalism ad-
mitted in a recent article, "Many psychologists who are inclined to accept
the functionalist framework are nevertheless worried about the failure of
functionalism to reveal much about consciousness. Functionalists have
made a few ingenious attempts to talk themselves and their colleagues out
7
of this worry, but they have not, in my view, done so with much success."
Let us therefore consider a number of theories in which consciousness is
granted a real existence. We will begin by examining the identity theory,
dual-aspect theories, and the theory of psychophysical parallelism. These
theories have many subtle philosophical nuances, but they all share one
key feature that is ruled out by the analysis presented in this chapter. We
can show this by briefly defining these theories:
8
(a) Identity theory. According to the identity theory, both conscious
mental experience and physical phenomena are real. However, mental
and neural events are one and the same, and they are basically physical.
(b) Dual-aspect theories. Perhaps the most well-known of these

9
(heories was devised by Spinoza, who held that there is one underlying
substance that has both physical and psychological aspects. Another
theory, devised by Alfred N. Whitehead, features one fundamental pro-
cess that is inherently endowed with "feeling," and that generates both
10
consciousness and physical phenomena. Many of these theories hold
that all physical objects are to some extent sentient, a position known
as panpsychism.
11
(c) Psychophysical parallelism. This theory holds that consciousness
and material phenomena are real, and are correlated with one another in
a one-to-one fashion without any causal connection. John von Neumann
jja ve the following interesting definition of psychophysical parallelism: " I t
is a fundamental requirement of the scientific viewpoint—the so-called
principle of the psycho-physical parallelism—that it must be possible so to
describe the extra-physical process of the subjective perception as if it
were in reality in the physical world—i.e. to assign to its parts equivalent
12
physical processes in the objective environment, in ordinary space."
The common feature of these theories is that they all posit a direct, one-
to-one correlation between the contents of consciousness and material
phenomena. They maintain that these two things are identical, or that they
are aspects of a third thing, or that they somehow run in parallel. Yet we
have seen that there can be no one-to-one correlation between the contents
of consciousness and the physical phenomena of the brain. Rather, the
correlation that must exist between these two real phenomena is highly
complex, and possesses the indirect character indicated by statements (1)
and (2).
If we posit a one-to-one correlation between the contents of conscious-
ness and certain natural phenomena, then these phenomena must be dis-
tinct from the neural interactions of the brain. This conclusion certainly
contradicts theories (a) and (c), which are intended to show that con-
sciousness can be comfortably accommodated within the framework of
current scientific thinking. This conclusion also contradicts those theories
of type (b) which hold that the underlying "substance" is co-extensive with
matter. (It is interesting to note that Whitehead's process philosophy does
not maintain this, and allows for the existence of extraneural psychic pro-
cesses that transmit information back and forth between the brain and an
13
extraneural sentient self. )
14
Another theory is idealism, which holds that only conscious minds
have actual existence, and that physical objects are nothing but mental
perceptions. This implies that the neurons in our brains must also be per-
ceptions, and the question arises, " I n what mind are they perceptions?" On
the one hand, it is hard to see how the neurons in my brain could be per-
ceptions in my own mind, for I have no direct awareness of them. On the
other hand, if these neurons are perceptions in some other mind, then the
individual conscious self must be distinct from the brain.
Let us go on to consider two other theories called interactionism and
epiphenomenalism. According to these theories, the conscious self is an
entity that is distinct from the body, and that can be influenced by physical
phenomena occurring in the brain. The difference between these theories
is that interactionism allows consciousness to influence the body, whereas
epiphenomenalism does not permit this. According to interactionism, the
conscious self receives sense impressions from the neural apparatus of
the brain, and is able to exert its will on the body by inducing changes in
neural activity. In epiphenomenalism the conscious self is simply a passive
spectator of events that are conducted entirely by physical processes.
Both interactionism and epiphenomenalism are compatible with the
analysis given in this chapter, for we have only considered the transmission
of information from the bodily machinery to consciousness. We should
note, however, that this process of information transmission, as we have
The Conscious Self as a Complete Sentient Personality 41
understood it, is quite different from the processes that many proponents
of epiphenomenalism seem to envision. Advocates of this theory often
convey the impression that consciousness is directly produced by the brain,
as we see in the statement that "the brain secretes thought, just as the liver
15
secretes bile." They suggest that matter in appropriate states of complex
organization automatically generates conscious awareness reflecting that
state of organization.
We have observed, however, that the contents of consciousness cannot
correspond directly with the material configurations of either a brain or a
hypothetical sentient machine. Rather, as we saw in Figures 3 and 4, the
contents of consciousness can only correspond in a one-to-one fashion with
higher-order abstract properties of these configurations. Now, why should
it happen that certain highly abstract features of complex material systems
should come to be represented by real, but nonphysical, conscious percep
tions? It is certainly misleading to compare such a process of representation
with the secretion of bile by the liver.
As we have pointed out in statements (1) and (2), some entity, process,
or law is required that can recognize the appropriate abstract patterns, and
modify the contents of consciousness accordingly. This entity or process
must be nonphysical or transphysical itself, for otherwise how can it either
influence or generate something nonphysical? Furthermore, if it is to be
capable of recognizing abstract patterns, this entity or process must be
endowed with powers of discrimination of the kind we associate with
intelligence.
2.5 The Conscious Self as a Complete Sentient Personality

One natural way of understanding statements (1) and (2) is provided
by the model of the conscious self that we introduced in Chapter 1. There
we drew upon the Bhagavad-gītā to formulate a picture of the conscious
self as an independent, irreducible being endowed with its own inherent *
senses. Here we propose that the entity or process described by statements
(1) and (2) corresponds with the innate senses or sensory processes of the
conscious self.
We are suggesting that the conscious self, or jivātmā, is a complete
sentient personality that is capable of riding in a material body just as a
passenger can ride in a vehicle. As such, each jīvātmā is conscious and
possesses all the attributes of a person, including senses and intelligence.
Just as a passenger can obtain information about his vehicle and its sur
roundings by interpreting certain instruments within the vehicle, so the
jīvātmā can ascertain the conditions in its body and bodily environment by
interpreting the physical states of the body's brain.
We realize, of course, that many other interpretations can be offered
for statements (1) and (2). Yet, most of these seem more complicated and
obscure than the simple hypothesis we are presenting here. (Sec Chapter 4
for a discussion of general requirements for a natural law or process that
can generate consciousness.) By presenting the body as a vehicle for the
jīvātmā, this hypothesis enables us to readily understand why nonphysical
consciousness should be associated with complex automata. In addition,
the hypothesis opens up the possibility that we may be able to enlarge our
understanding of the conscious self by direct sensory experience. We shall
therefore briefly discuss this hypothesis and consider some objections that
might be raised against it.
The position of the jīvātmā as the conscious perceiver of the body can be
illustrated by the example of a person reading a book. When a person
reads, he becomes conscious of various thoughts and ideas corresponding
to higher-order abstract properties of the arrangement of ink on the pages.
Yet none of these abstract properties actually exist in the book itself, nor
would we imagine that the book is conscious of the story it records. As we
have indicated in Figure 5, the establishment of a correlation between the
book and conscious awareness of its contents requires the existence of a
conscious person with intelligence and senses who can read the book. Simi
larly, we can most readily understand the relation between consciousness
and certain abstract properties of brain states by postulating the existence
of an intelligent, sentient entity that can read these states.
Now, the objection will be raised that if we try to explain a conscious per
son by positing the existence of another conscious person within his body,
then we have actually explained nothing at all. One can then ask how the
consciousness of this person is to be explained, and this leads to an infinite
regress.
This objection presupposes that an explanation of consciousness must be
mechanistic. But we must simply face the fact that conscious personality
cannot be explained mechanistically. An infinite regress of this kind is in
fact unavoidable—unless we either give up the effort to understand con
sciousness or posit the existence of a sentient entity that cannot be reduced
to a combination of insentient parts.
The reductionistic approach to explanation will not be fruitful when
applied to the conscious self. We will never be able to explain conscious-
' ness by breaking it down into combinations of simple insensate compo
nents. Nor can it be satisfactory to "explain" consciousness by sticking it
Figure 5. The relation between consciousness and the physical structures of a book.
When a person reads a book he becomes aware of higher-order abstract properties
of the patterns of ink on paper—properties that are not directly present in these
physical structures. This observation suggests an answer to the questions posed in
the captions to Figures 3 and 4. We can answer these questions by positing the exis-
tence of a nonphysical sentient agency that can read the abstract features of the
computer or brain. In the case of the computer there may be no subjective experi-
ence, and no need to postulate such an entity. But since the existence of subjective
experience in humans is indubitable, we cannot avoid the conclusion that some
entity must exist that is capable of reading the physical states of the brain.
awkwardly into a mechanistic world picture as an inexplicable, incompa

tible element. We can, however, seek to discover the true potential of the
conscious self, and the nature of its relation with reality as a whole. If the
conscious self, or jīvātmā, is indeed an independent sentient entity riding in
a material vehicle, then the possibility arises that the jīvātmā may be able
to occupy different kinds of bodies, or function without any material body
at all. If the sensory input of the embodied jīvātmā is being filtered through
a material data processing system, then it is possible that the jīvātmā
may be able to experience a more direct and vivid mode of sensory percep
tion on a higher level of consciousness. If such possibilities could be real
ized by direct experience—using practical, reliable methods—then we
would indeed be able to feel that we had increased our understanding of
consciousness.
In Chapter 9 we will discuss how such a direct investigation of conscious
ness might actually be carried out. Here we will conclude by making a few
observations about the relation between our model of the conscious self
and the theories of interactionism and epiphenomenalism. As we have
observed, the arguments presented in this chapter do not distinguish be
tween these two theories. However, for the sake of completeness we
should point out that our model of the self must be classed under the head
ing of interactionism. According to the Bhagavad-gītā, the jīvātmā does in
fluence the functioning of the material body, but this influence is exerted in
an extremely subtle way. We will discuss this in some detail in Chapter 7.
For now we will briefly mention some of the standard objections to
interactionism and show that they do not apply to our model.
In Western philosophy the most famous version of interactionism is that
of Rene Descartes, who proposed that human beings (but not animals)
possess a noncorporeal mind that directs the actions of their bodies. Ac
cording to Descartes, matter has the properties of extension and location
in space, but mind does not. Consequently, many people have rejected
Descartes' theory on the basis that it is very hard to see how an entity with
no spatial properties could possibly interact with something located in a
particular position in space. 16
This objection to interactionism certainly makes sense, but it only ap
plies to Descartes' particular theory. It can easily be answered by noting
that a nonphysical entity does not have to be devoid of all physical proper
ties. We have tried to formally define what is meant by "nonphysical" by
introducing the idea of numerical indescribability. We say that an entity is
numerically indescribable if significant features of the entity cannot be
represented by numbers. Yet this does not imply that the entity cannot
have any measurable properties at all.
Consciousness provides our archetypal example of numerical inde-

scribability. Yet according to the Vedic literature, the conscious self is
localized within the body. For example, the Mundaka Upanisad states:
eso 'nurātmā cetasā veditavyo

yasmin prānah pañcadhā samviveśa
prānaiś cittam sarvam otam prajānām
yasmin viśuddhe vibhavaty eśa ātmā
"The conscious self is atomic in size and can be perceived by perfect intelli
gence. This atomic self is floating in the five kinds of air [prāna, apāna,
vyāna, samāna, and udāna], is situated within the heart, and spreads its
influence all over the body of the embodied living being. When the self
is purified from the contamination of the five kinds of air, its spiritual
influence is exhibited." 17
We note that the jīvātmā is said to be extremely minute in size, and to be
located in the region of the heart (rather than in the brain). However, the
jīvātmā does not interact directly with the gross physical structures of the
body. Rather, it interacts with subtle material elements, and these in turn
interact with gross matter in accordance with principles that have yet to be
discovered by present-day physicists and chemists.
This brings us to another objection that is commonly raised against the
theory of interactionism—the objection that the laws of physics must be
violated if a nonphysical conscious entity is able to influence the behavior
of matter. We can respond to this objection by pointing out that because of
computational difficulties, it is nearly impossible to determine whether or
not a complex system really does obey the known laws of physics. Using
quantum mechanics, we have great difficulty analyzing a single molecule of
benzene, and we cannot even begin to predict what will happen in a brain.
Furthermore, developments within the science of physics have led to re
peated revisions in the accepted laws of nature, and we may expect to see
similar revisions in the future. We can only conclude that we have no justifi
cation for rejecting interactionism simply because it contradicts the alleged
universality of a given system of physical laws.
As a final point, we note that the physicist Eugene Wigner has invoked
the principles of current physical theory to argue in favor of an essentially
interactive model of consciousness. Wigner argues that in physics, causa
tion is never a one-way affair. Thus, if consciousness is real, as Wigner
believes, then consciousness must exert some influence on matter. Con
sequently, Wigner concludes that "the present laws of physics are at least
incomplete without a translation into terms of mental phenomena. More
likely they are inaccurate, the inaccuracy increasing with the increase in the
18
role which life plays in the phenomena considered."
Notes
1. M.I.T. college catalogue, 1979, p. 118.
2. In actual computer applications much more sophisticated methods of
calculating square roots would be used. The method presented in
Figure 1 is intended to provide a simple example of the nature of
computer programs.
3. Kleene, Introduction to Metamathematics, pp. 317-322, 377-381.

4. Winston, Artificial Intelligence, p. 253.
5. For the sake of clarity, let us briefly indicate why this is so. Suppose
that a model of a sentient entity can be described by means of a com-
puter program. Then a certain level of organization of the program will
correspond to the elementary constituents of the model. For example,
in a quantum mechanical model these constituents might be quantum
wave functions. The level of the program corresponding to "thoughts"
and "feelings" will be much higher than this level. Hence this "cog-
nitive" level will not in any sense exist in the actual system being
modeled. It will correspond only to abstract properties of the states of
this system, and thus "something" of the kind described in (1) and (2)
will be needed to establish the association between the system and the
contents of consciousness.
6. Fodor, "The Mind-Body Problem," pp. 114-123.

7. Fodor, p. 122.
8. Edwards, ed., The Encyclopedia of Philosophy, Vol. 5, p. 339.

9. Edwards, p. 340.
10. Whitehead, Process and Reality.
11. Edwards, p. 342.
12. von Neumann, Mathematical Foundations of Quantum Mechanics,
pp. 418-419.
13. In Whitehead, p. 339, we read that the ultimate perception of bodily
sensations takes place in a sentient process operating in "empty space
amid the interstices of the brain."
Notes 47

16. Edwards, p. 341. This point is also raised in Fodor, p. 114.
17. Mundaka Upanisad, 3.1.9. This verse is cited in A.C. Bhaktivedanta
Swami Prabhupāda, Bhagavad-gītā As It Is, p. 95.
18. Wigner, "Physics and the Explanation of Life," p. 44.
Chapter 3
Dialogue
On Consciousness
And
The Quantum
A Conversation Between
Dr. Felix Avaroha, Mathematician
Dr. James Yantry, Biologist
Dr. Hans Kutark, Physicist
Who Are Later Joined by Two Physicists,
Dr. Sophus Baum
Dr. Francesco Shunya
The Time: Summer, 1980

The Place: Boston, Massachusetts
This dialogue deals with the question of whether or not modern physics
can provide an adequate description of conscious life. The characters are fic
titious, but most of the views they express have been expounded by various
scientists and philosophers. Sections 3.1, 3.2, and 3.3 are devoted to a dis
cussion of consciousness and the problems of quantum epistemology. In
Section 3.4 the concept of the jīvātmā, or nonphysical conscious self, is in
troduced, and there is a discussion of how the nonphysical self could be
investigated experimentally.
Avaroha: Over the past three hundred years there has been considerable
advancement in many scientific fields, but our understanding of the
nature of consciousness has remained in a completely undeveloped
state. This imbalance in our scientific world view has resulted in a
highly distorted conception of the nature of the conscious self. This
conception is rendered particularly unsatisfactory by the widespread
tendency to attribute universal validity to scientific hypotheses, and to
regard modern science as the only source of true knowledge about the
world.
I propose that we need a science of the conscious self, and that such a
49
50 Dialogue on Consciousness mul thr <Jmiinitm
science requires a source of valid and substantial data directly pertain-

ing to the nature of consciousness. Since such data in lacking in all the
fields of modern science, from physics to psychology, I propose that
we should turn to ancient and reliable sources of knowledge that we
have hitherto neglected in our scientific studies. One such source is the
Bhagavad-gītā, which presents both a fundamental conceptual frame-
work for the understanding of consciousness, and piuclical means lor
directly studying the conscious self.1
If the limited and imperfect nature of our current scientific theories
is recognized, it will be possible to harmonize the valid insights of mod-
ern science with the broader knowledge contained in the Bhagavad-
gītā. Not only will this provide us with a much-needed advance in our
understanding of the nature of conscious life, but also it will yield many
stimulating ideas relevant to our existing studies of matter and its
interactions.
Yantry: I don't feel that the program you have outlined is at all justified.
There are no serious deficiencies in our current scientific world view,
especially in my own field of biology. True, there are many mysteries in
the science of life, and these are the subject of the most vigorous inves
tigation at the present time. But the basic principles that form the foun
dation of our scientific research are solidly established. The edifice of
scientific progress is firmly founded on the assumption that all phe
nomena in the universe are due to certain basic physical processes.
Therefore, life too must be a result of physical processes only, and the
course of life must be determined automatically by the physical and
chemical occurrences within living matter.'
Although there may be many unsolved problems in high energy
physics, these problems all involve extreme conditions that do not
arise in living systems. The basis of life is chemistry, and all the phe
nomena of chemistry are now completely understood in terms of
3
the physics of atomic and molecular interactions. A l l the elements
that make up living systems are known, and the laws governing them
are also known. You seem to be advocating the introduction of some
kind of nonphysical principles or entities into our scientific picture
of life. Since the physical principles underlying life are completely
known, there is quite simply no room in our scientific world view for
such vitalistic ideas. 4
Avaroha: It is well known that our experience of conscious perception can
not be described within the mechanistic framework you are advocating
as the sole basis for the science of life. 5 As a standard example, con-
Dialogue on Consciousness and the Quantum 51
sider our conscious perception of the color red. You can specify the
wavelength of red light in quantitative terms, and you can also describe
how that light induces certain chemical reactions in the retina of the
eye. You can go o n , at least in principle, to describe a great variety of
consequent physical and chemical transformations in the nerve cells of
the retina and the brain. But at no time in the course of this description
do you say anything at all about the perception of red itself. This per-
ception takes place and is therefore certainly real, but our scientific
theories do not enable us to say anything about it.
Yantry: This problem of consciousness has been discussed at great length
6
by many philosophers, but I don't think it has any bearing on our
theories of physics and chemistry. The general consensus is that con-
sciousness must be generated by matter when it reaches certain levels
of complex organization, as in the brain. Some speak of consciousness
as something distinct from matter, and refer to it as an "epiphe-
nomenon." Others speak of it as an aspect of matter. In any event, the
important point is that consciousness does not exert any influence on
7
the behavior of matter in living systems. This behavior is entirely
determined by the known laws of chemistry and physics.
Avaroha: When consciousness is treated in this way, it occupies a very
awkward and artificial position in our theoretical picture. In the sci-
ence of physics, all other products or aspects of a physical system have
some effect on the behavior of the system as a whole. Why should
8
consciousness be an exception?
I suggest that your characterization of consciousness is inadequate,
and that a valid account must describe how consciousness interacts
with matter. To clear the way for this, I would like to point out that the
physical principles underlying the science of chemistry are not fully
understood. There are serious problems in the theory of quantum
mechanics, which describes the behavior of atoms and molecules.
These problems have been the subject of great controversy, and they
9
have resulted in a great tangle of speculative interpretations, some of
which go to desperate extremes. It is clear that some modification of
the theory is necessary. I would suggest that in making such modifica-
tions, we should keep in mind the requirements for an adequate theory
10
of consciousness.
Yantry: I am not an expert in the theory of quantum mechanics, but I
doubt very much that any of the problems in that field could have any
11
bearing on our established physiochemical understanding of l i f e .
Perhaps, Kutark, you could shed some light on this matter.
52 Dialogue on C Onsiiousnc.w ami ilif Quantum
Kutark: The situation is perhaps more complicated than you seem to

think, Yantry. Yet I am convinced that the quantum theory, as it
stands, provides a complete and satisfactory account of atomic inter-
12
actions. In my opinion, Avaroha, the problems you mentioned can
be resolved by a proper understanding of the subtle features of quan-
tum epistemology. But if you like we can discuss them, and if your view
prevails we can go on to consider your more radical ideas. What do you
think, Yantry?
Yantry: I'm surprised you would even consider the idea that our funda-
mental scientific theories might be in need of significant revision. Well,
go ahead. I am sure that your final conclusions, if they are correct, will
be consistent with the basic viewpoint on which all our work is based. I
simply ask that you try to keep the discussion accessible to persons who
are not specialists in the field of physics.
3.1 A Quantum Mechanical Problem

Avaroha: Let us proceed, then, by looking at a simple example. Consider
the Wilson cloud chamber. As you know, this device consists of a glass
chamber containing a piece of radioactive substance. If the humidity in
the chamber is maintained just at the point of saturation, an observer
will see thin lines of fog emanating from the radioactive material. Can
you explain this?
Yantry: Of course. This is an elementary experiment used in beginning
classes to demonstrate the phenomenon of radioactivity. As the radio-
active atoms in the chamber decay, they emit subatomic particles that
travel at very high velocities. These particles ionize the air along their
paths, and water droplets condense around the ions, marking the paths
with visible lines of fog.
Avaroha: Presumably, then, the established laws of physics should be
able, at least in principle, to give a complete account of a man in a room
watching a cloud chamber.
Yantry: Yes. Of course, we should keep in mind that modern physics is not
deterministic. An element of chance must enter into the physical de-
scription, but this is unavoidable. We now know that chance is part of
the nature of things.
Avaroha: Could you give an example of this?
Yantry: You have already mentioned one of the standard examples. The
time when a given radioactive atom decays cannot be specified with
certainty, even in principle. We can specify at most the probability that
the atom will decay within a given interval of time. Likewise, the
A Quantum Mechanical Problem 53
direction in which the high energy particle is emitted is completely

random.
Avaroha: Nonetheless, you do agree, don't you, that a given atom does
decay and produce a definite visible track in a particular direction at
some particular time?
Yantry: Yes, of course.
Avaroha: Do you also agree that once this happens it will be perceived
by the man as a definite event, and that the laws of physics should en-
able us to give a complete description of the process whereby this
perception takes place?
Yantry: This is certainly true.
Avaroha: Very well. For simplicity, let's assume that a single radioactive
atom is fixed in the center of the cloud chamber. We expect to find that as
Figure 1. Cloud c h a m b e r track produced by the decay of a radioactive a t o m . T h e

track is believed to have been p r o d u c e d by an alpha particle emitted by an atom of
radon situated to the left of the picture. ( F r o m C . T . R . Wilson, Proc. Roy. Soc.
London (A) 8 7 , 2 7 7 (1912).)
the man watches, he may see nothing at all for some time. Then, at some
particular moment, he will see a single trail of fog extending in a straight
line from the center of the chamber. Let us begin by considering what
happens within the atom to produce this trail.
Kutark: We can easily explain this in terms of the quantum mechanical
"tunneling" effect. We can imagine that the particle is trapped within the
nucleus of the atom by a potential barrier so high that the particle lacks
sufficient energy to escape. In quantum physics such a particle has a
certain chance of passing through the barrier, even though this is im-
possible from the point of view of classical physics. This phenomenon
is called tunneling, and it is the standard explanation for radioactivity.
Avaroha: This kind of verbal account is often given, but quantum mechani-
cal tunneling is not actually described in terms of a localized particle.
Rather, it is described in terms of a wave. In this case we can visualize
the potential barrier that you mentioned as a partially reflecting glass
54 Dialogue on Consciousness and the Quantum
shell. If some light were trapped within the shell, it would bounce
back and forth repeatedly, but with each reflection some light would
pass through the shell. Similarly, the process of radioactive decay is de-
scribed in terms of a wave that is partially trapped within the nucleus,
and steadily leaks out, spreading uniformly in all directions.
Kutark: It is true that the word "particle" is used in a metaphorical sense in
the quantum theory. Physical descriptions are actually given in terms
of waves, or more generally, in terms of "state vectors" in Hilbert
space, which can be represented mathematically in various ways.
Yantry: I take it that this talk about waves and particles must be related to
Figure 2. T h e q u a n t u m mechanical wave describing the emission of a particle from

the nucleus of a radioactive atom.
What Quantum Mechanics is Really Saying 55
the "wave-particle duality" that we often hear about. But your state
ments seem rather confusing. You say that the quantum theory actu
ally describes radioactive decay in terms of a wave, and that the word
"particle" is used metaphorically. Yet I can understand how a particle
can produce a definite track, whereas it is hard to see how a wave that
spreads uniformly in all directions can do so.
Kutark: The answer is very simple. Thus far, we have considered only the
radioactive "particle" itself. To properly describe the cloud chamber
tracks, you would have to extend the physical system at least to the
point of including the atoms of air within the chamber. You would also
have to take into account the interaction between these atoms and the
particle. If you do this, you will find that the state vector of the system
provides a completely satisfactory account of the origin of the tracks.
But in this situation we cannot really call the state vector a wave; it is a
much more complicated mathematical construct.
3.2 What Quantum Mechanics is Really Saying

Avaroha: Yes. To really understand the quantum mechanical treatment of
our cloud chamber experiment, we will have to deal with much more
complicated ideas than we have considered thus far. Let us therefore
enlarge our physical system as you have suggested and introduce a gen
eral quantum mechanical state vector, Ψ(t), that describes its physical
state. In addition, we must introduce the Schrodinger equation,
which specifies how the state of the system changes with the passage of
time. In this equation, the Hamiltonian operator, H, sums up all the
physical laws of cause and effect operating within the system.
Yantry: Naturally, I have seen the Schrodinger equation before, but I have
not studied it in detail. It involves very complicated calculations that
only a specialist can understand. Isn't it possible to make your points
about quantum mechanics using simple examples that can be readily
visualized?
Avaroha: Superficial verbal accounts of quantum mechanics are generally
quite misleading, and may even completely misrepresent the theory.
To really appreciate what is going on in the quantum theory, we must
obtain some conception of how the theory actually describes nature.
In this case, however, it should be possible to present things in a
fairly nontechnical way if we accept one basic mathematical transfor

mation of the Schrodinger equation. We begin by dividing the total sys
tem into two parts: the radioactive particle itself, and everything in the
system except for this particle. I shall refer to the total system as Sys
tem I, and I shall refer to the total system minus the radioactive particle
as System II.
The state vector Ψ(t) can be expressed as,
where the Fk(t)'s are state vectors for the radioactive particle itself,
and the Xk(t)'s are state vectors for the rest of the system. If we choose
the X k (t)'s so as to form what is called a " c o m p l e t e o r t h o n o r m a l basis,"
then we can express any Ψ(t) in this way in terms of some suitable set of
Fk(t)'s.
Kutark: You are simply describing the standard m e t h o d of solving differ
ential equations called separation of variables.
Avaroha: That is correct. F o r o u r purposes, however, it is sufficient to
point out that in the q u a n t u m theory the mathematical entities Fk(t)
and Xk(t) for k=1, 2, 3, . . . completely define the state vector of o u r
total physical system at time t.
Yantry: I'm afraid this all seems very abstract. Could you give me some
idea of what these "state vectors" really are?
Avaroha: E a c h Fk(t) refers to the radioactive particle and can be repre
sented as a wave propagating through three-dimensional space. Since
Ψ(t) and the X k (t)'s refer to very complicated systems involving vast
numbers of particles, they cannot be represented in such a readily visu-
alizable form. We will assume that the X k (t)'s are solutions of the
Schr dinger equation for System I I , considered by itself. With this as
sumption you can think of the Xk(t)'s as q u a n t u m mechanical histories
of what will h a p p e n in the system in the absence of the radioactive par
ticle. Likewise, you can think of Ψ(t) as a history depicting what will
happen if the particle is present.
Yantry: Could you explain this in m o r e specific terms?
Avaroha: Yes. Consider this example: Suppose that System II consists of n
molecules of gas located at fixed positions in the cloud c h a m b e r . 1 3 You
can think of these molecules as potential targets for the radioactive
particle. T h e following drawing shows how you might visualize this
situation in terms of classical physical concepts. [See Figure 3.] H e r e
the particle is depicted as a small sphere that has emerged from the
Figure 3. The process of radioactive decay as it might be visualized from the view-
point of classical physics.
radioactive nucleus at a high velocity and is in the process of colliding

with three of the target molecules.
Unfortunately, although this diagram is readily understandable, it
completely misrepresents the process of radioactive decay as described
in the quantum theory. To present the q u a n t u m mechanical descrip-
tion, I will assume, for simplicity, that System II consists of n idealized
target atoms, each of which can be either excited (e) or unexcited (u).
Since these atoms interact only with the radioactive "particle" and not
with one another, System II remains unchanged with the passage of
n
time. Thus each history, X (t), of System II is static, and there are 2
k
58 Dialogue on Consciousne.w ami llw (Juunium
possible X (t)'s, one for each possible arrangement of executed and un-
k
excited states of the n atoms. In this setting the behavior of the total
system of target atoms plus "particle" can be depicted us shown in
Figure 4. Here nine target atoms are shown in the smile positions as in
Figure 4. Part of the state vector for a system of nine target atoms and a radioactive
nucleus. Each target atom can be in an excited state (e) or an unexcited state (u).
The state vector is represented by waves F , F , F , .......,F , corresponding to
0 1 2 511
the 512 possible combinations of excited and unexcited states of the nine atoms.
Each box depicts a combination of excited and unexcited states, and the corre-
sponding wave. These waves can " f l o w " from one box to another if the two boxes
differ in the state of excitation of one atom, and the waves and atoms are lined up
properly. (For mathematical details see Note 24.)
Figure 3. The circles marked with eor u represent the nine atoms, and
9
four of the 2 =512 possible patterns of excitation are shown.
Corresponding to each X (t), or pattern of excitation, there is an
k
F (t) which is depicted as a wave flowing through space. The wave

k
F (t) flows radially out of the radioactive nucleus, and is similar to the
0
wave shown in Figure 2. This wave generates respective waves F (t), 1
F (t), and F (t), which are associated with X , X , and X , as shown in

2 3 1 2 3
the figure. The general principle is that F can generate a wave F if the i j
corresponding X differs from X in the state of excitation of a single

i j
atom, and this atom lies in the path of F . The figure shows how the F- i
waves, by propagating in this way, can start from X (t), where there 0
are no excited atoms, and eventually reach X (t), where there is a line 3
of three excited atoms extending out from the radioactive nucleus.

Kutark: This shows in a pictorial way how a trail of excited atoms is
formed. If X corresponds to a chain of m excited atoms extending in a
i
straight line from the nucleus, and X corresponds to the same chain j
with one additional excited atom at the end, then the F-waves can pass
from X to X if this atom is lined up with the chain. Otherwise they
i j
cannot do this, for the waves tend to propagate in a straight line.

We can also see why a particle of lower energy will tend to produce a
crooked track: At lower energies the F-waves will have longer wave-
lengths, and they will tend to spread put more. Thus the F-waves will
be able to reach X s corresponding to curved tracks of various shapes.
k
Yantry: This seems very obscure. Why, if we are dealing with one radio-
active particle, are you considering many different F-waves propagat-
ing through three-dimensional space? Do your successive waves F , 0
F , F , and F correspond to successive stages in time in the formation

1 2 3
of the chain of excited atoms? Also, out of all possible straight chains
of atoms extending from the nucleus, how is it determined which one
will be traversed by these waves?
Avaroha: It is actually incorrect to think of the F-waves as real phenomena
occurring in three-dimensional space. Rather, each F is an abstract k
mathematical function associated with a possible arrangement of ex-

cited and unexcited atoms in the cloud chamber. Rather than flowing
through three-dimensional space, the F-waves are flowing through an
n
abstract realm of possibilities. In principle, any of the 2 different pat-
terns of atomic excitation are possible, but due to the way the F-waves
propagate, only those patterns representing straight tracks are as-
sociated with a non-zero F . Nothing has been said, however, about
k
choosing a particular direction of propagation. Tracks in all possible

directions are equally and simultaneously represented in the total state

vector, Ψ(t), that we are trying to describe.
I should also point out that the figure is referring to one moment in
time. The successive Xk's and Fk's do not refer to successive steps in
the propagation of the "particle." In fact, the Fk(f)'s remain practically
unchanged during the period of perhaps an hour or so, in which the
observer is watching the cloud chamber. This is because the intensity
of the F 0 wave emanating from the radioactive nucleus is diminished by
one half during a period of time equal to the "half-life" of that radio
active element. If this is substantially longer than the period of ob
servation, we can assume that Ψ(t) remains nearly static during this
time.
Yantry: That means that this complicated Ψ(t) specifies neither the time of
formation of the cloud chamber track, nor its direction. Doesn't this
mean that Ψ(t) must refer to a set of alternative events from which the
real event is chosen by some stochastic process? What do you mean,
then, by saying that Ψ(t) refers to the physical state of the system at a
particular moment in time?
Avaroha: That is an interesting question. As far as the quantum theory is
concerned, the state vector for a physical system provides a complete
description of the state of that system, in the sense that the system actu
ally possesses no physically relevant properties or characteristics not
specified by the state vector. Isn't this so?
Kutark: That is true. We would be contradicting the quantum theory if we
were to attribute to a physical system any property not specified by its
state vector. 1 4
3.3 Threading the Labyrinth of Quantum Epistemology

Avaroha: So, we seem to have a situation in which the state of our physical
system is not characterized by any particular time of appearance of the
track, or any particular direction of propagation. How do we reconcile
this with the fact that, at a particular time, our observer witnesses the
appearance of one particular track?
Kutark: The answer to this question involves an analysis of the quantum
mechanical process of measurement. The state vector Ψ(t) refers to
phenomena that are atomic, and not directly observable. To be ob
served, they must produce effects that are amplified to a macroscopic
level by some suitable apparatus, in this case the cloud chamber. Nor
mally, certain features of the state vector will correspond with certain
Threading the Labyrinth of Quantum Epistemology 61
gross features of the measuring apparatus—features that can be di-

rectly observed and then described in the language of classical physics.
In this case the different X 's representing straight tracks correspond
k
to different visible trails of fog within the chamber. Here, amplifica-

tion of the quantum phenomena is due to the process of condensation,
whereby droplets of water coalesce about the ionized atoms.
At the time of the measurement, we say that the apparatus will be
in the macroscopic state corresponding to X , with a probability of
k
(F ,F ). Here (F ,F ) represents what we could call the total weight

k k k k
of F , and it is interpreted as a probability. If X represents either a

k k
straight track or no track at all, then, as we have seen, F will have k
some positive weight. Thus these X s correspond to possible observ-

k
able alternatives. In contrast, if X represents some highly irregular

k
arrangement of excited atoms, then F will be zero, and we will not

k
expect to observe a track corresponding to this X . Is this satisfactory?

k
Avaroha: Are you trying to say that the cloud chamber must be described
in terms of classical physics?
Kutark: Yes. Niels Bohr pointed out that the quantum theory is meaning-
less unless it can be translated into the language of classical physics,
15
which we use to describe our observations of the w o r l d . Therefore,
in the quantum mechanical process of measurement, we must change
from the quantum mechanical to the classical mode of description at
some point. In this case it is natural to do this at the point where the ion
trails, which we describe quantum mechanically, give rise to visible
vapor trails within the cloud chamber.
Avaroha: Is classical physics capable of giving an adequate account of the
structure of matter?
Kutark: No. As you well know, the quantum theory was developed be-
cause classical physics could not successfully describe the atomic
constituents of matter.
Avaroha: Are you saying, then, that it is not possible for us to give an
16
adequate description of the cloud chamber?
Kutark: Not necessarily. We could describe the cloud chamber by means
of quantum mechanics and revert to classical physics at another point
in the process of observation. For example, we could do this at the
point where light reflecting from the cloud chamber was focused on the
retina of the observer's eye. John von Neumann pointed out that the
boundary between the observer and the observed system can be arbi-
trarily shifted, with everything on the observed side being completely
17
describable by quantum mechanics.
Avaroha: This procedure leaves the conscious observer on the side of the
boundary where phenomena are to be described in terms of in
adequate classical concepts. If you a r e truly forced to divide the world
into two parts in this way, then it would seem that the quantum theory
is incapable of adequately describing the observer. F u r t h e r m o r e , since
your boundary singles out one potential observer for unique treat
ment, you seem to be introducing an element of solipsism into your
theory. If the quantum theory simply describes one individual's sub
jective impressions of the part of the world external to h i m , 1 8 then it
can provide neither a complete description of reality nor an adequate
account of consciousness.
Kutark: Actually, the observer is irrelevant. Once the observer is brought
into the picture the impression is created that a subjective element has
been introduced into physics. This simply entangles us in muddled
thinking! To maintain scientific objectivity we should exclude all con
scious observers from the physical system and define the measuring!
process in terms of some automatic recording device describable in
terms of familiar, classical physical c o n c e p t s . 1 9
Avaroha: What do we do if we want to give a theoretical account of the ac
tual human observer of our cloud chamber? Are you saying this can't
be d o n e ?
Kutark: Of course it can be d o n e . You can use one comprehensive state
vector to describe the cloud chamber, the human observer, and as
much of his environment as you may wish to include. We should simply
keep in mind that the state vector must be interpreted as an objective
description of physical reality.
Avaroha: Very well, then, let's try to do this. Why don't we adopt our pre
vious notation to this enlarged system? We can refer to the enlarged
system as System I, and let Ψ(t) designate its state vector. Similarly,
we can define System II as the enlarged system minus the radioactive
particle. T h e F k (t)'s and Xk(t)'s can then be defined as before.
If we adopt these conventions, we can view the X k (t)'s as possible
histories of the total system minus the radioactive particle. Most of
these histories will presumably represent matter in various chaotic
states. But if quantum mechanics can actually describe System II
adequately, then some of these histories must represent the formation
of a vapor trail in the cloud chamber, followed by the perception of this
trail by the observer.
Kutark: Yes. For each possible track in the cloud chamber there will be an
X k that represents the formation of that track through condensation.
and its subsequent perception by the observer.

Avaroha: Can you use the Schrodinger equation to demonstrate rigor
ously that these X k 's really will adequately describe the physiological
processes going on in the body of the observer?
Kutark: Of course not. That would plainly be asking too much. We simply
assume this because if it were not so, then our theory would be wrong.
Avaroha: All right. Let's make this assumption for the sake of argument.
If we do this we will find, as before, that the total state vector Ψ(t)
corresponds to a superposition of Xk(t)'s representing either parti
cle tracks in various directions, or no track at all. Now, however, the
X k (t)'s also represent physical states of the observer corresponding to
his, perception of these conflicting alternatives. In other words, at a
gjven time t, the state vector Ψ(t) represents the observer as being
simultaneously in a large number of different physiological states cor
responding to mutually contradictory perceptions. If Ψ(t) is indeed a
complete description of the physical state of the system, then how do
you account for this?
Figure 5. The state vector for the total system, including the human observer. The
five images correspond to five situations of the observer, involving distinct experi
ences and perceptions. The state vector gives nearly equal representation to many
such situations.
(.4 Dialogue on Consciousness and the Quantum
Kutark: We deal with this situation by employing what is known as the "re
duction of the wave function." Essentially, I have already described
this procedure. At the time of observation the state vector Ψ(t) is
replaced by one of its unambiguous components, Fk(t)Xk(t), with a
probability of (Fk(t),Fk(t)). We can also express this by saying that at
the time of observation, Ψ(t) is replaced by the statistical mixture of
states.
Here we represent the statistical mixture by what is known as a von

Neumann density matrix. 2 0
Avaroha: Since Ψ(t) is intended to represent the physical state of the sys
tem, naturally we would suppose that changes in Ψ(t) must correspond
to changes occurring in nature in the physical system itself. In the
science of physics we normally suppose that natural transformations
of matter are due to certain physical forces. In quantum mechanics
all transformations of this kind are described by the Schr dinger
equation. I would therefore like to ask what phenomena in nature cor
respond to the abrupt "reduction" of the state vector that you are
introducing. Also, exactly what time are you referring to when you
speak of the "time of observation"?
Kutark: This subject has been plagued by a great deal of controversy, but
recently this controversy has been resolved in a very satisfying way. It
is quite true that the reduction of the state vector seems arbitrary and
nonphysical, especially because it can be defined in many different
ways. It has been shown, however, that at the time when the micro
scopic quantum phenomena are amplified to the point of producing
observable effects, a very important change occurs in the components
Fk(t)Xk(t) of Ψ(t). At this time these components lose the power to
interfere with one another, and thus they become effectively inde
pendent. Therefore, at this time our expression M(t) becomes equiva
lent for all practical purposes with Ψ(t), and can justifiably replace it.
Thus, the answer to your first question is that when the reduction of the
wave function is properly defined, it doesn't really represent a change
in our description of the physical system at all. In fact, this realization
constitutes the indispensible completion and natural crowning of the
21
magnificent edifice of quantum mechanics!
Avaroha: There seems to be some ambiguity in your statements. M(t) may
resemble Ψ(t) in certain respects. But it is formally different from Ψ(t)

and does not, for example, satisfy the Schr dinger equation. The most
striking similarity between M(t) and Ψ(t), however, is that both of
these mathematical entities fail to specify the specific alternative state
Fk(t)Xk(t) corresponding to the unique conscious experience of the ob
server. Therefore, the original problem we encountered with ^ ( f ) is
not solved if we simply replace Ψ(t) by M(t).
Moreover, you also said earlier that M(t) means that the system is in
a specific state Fk(t)Xk(t), and that this came about abruptly with a
probability of (Fk(t),Fk(t)). Thus you are using M(t) in two ways: to
describe the state of the system at time t, and to describe a process
whereby the system reached a completely different state by a random
jump. Which of these two conflicting usages would you like to choose?
You certainly cannot use both consistently.
Kutark: Many physicists have concluded that quantum mechanical state
ments must always refer to statistical ensembles. A l l we can say is that
out of N identically prepared systems, N(Fk(t),Fk(t)) will be in the
state Fk(t)Xk(t) at the time of observation. M(t) in fact expresses this
property of a large ensemble of systems. 22
Avaroha: Do you mean to say that in quantum mechanics we cannot refer
to the specific situation that is actually experienced by the observer at a
particular time? If this is so, doesn't it follow that quantum mechanics
is indeed drastically incomplete?
Kutark: Certainly not, for I have already said that you can refer to the
specific state of the observer by using a particular Xk(t).
Avaroha: You are saying then that the state of the system changes abruptly
from Ψ(t) to some particular Fk(t)Xk(t) at the time of observation?
Kutark: That is certainly one standard interpretation of quantum
mechanics. 23
Avaroha: I would like to refer once again to my question about the "time
of observation." You said that the time of observation is, approxi
mately, the time when the quantum phenomena are amplified from the
microscopic to the macroscopic level. But in our system, this process is
occurring continuously during the entire period of, say, an hour in
which the cloud chamber is being observed.
Kutark: Oh? I would think that the amplification must occur directly after
the chamber is triggered and the process of condensation can begin.
Avaroha: It is true that some cloud chambers are triggered at specific
times. But let's consider what happens if our cloud chamber is of
the type that operates continuously. In that case we will find that the
state vector Ψ(t) for our total system does not single out any particular
time of occurrence for the event of radioactive decay. Rather, W(t)
is a superposition of Fk(t)Xk(t)'s, representing all possible times of
observation for the track.
The structure of W(t) in this situation is rather interesting. As we
know, each Xk(t) can be thought of as a possible history of events in
System I I . In Ψ(t) each Xk(t) is present to a degree determined by the
magnitude of its coefficient, Fk(t). If Fk(t) is zero, then we can assume
that Xk(t) is not represented in Ψ(t).
I can show by a diagram how a history representing the observation
of a track becomes represented in W(t). [See Figure 6.] Here the his
tory X 0 represents what happens in System II when there is no ion trail
Figure 6. This diagram shows how a history Xn, representing the observation of a
particular track, is added to the total state vector. This is a process involving a
number of intermediate histories, Xk1,Xk2, . . . ,Xkm, that represent the observa
tion of partial tracks of greater and greater length. Here each history is indicated by
a line, and the line is made thicker during the period when that history is repre
sented in the total wave function. Vertical distance between histories designates
similarity, and we can see that a history can be added to the total state vector only at
a time when it is very similar to another history that is already represented there.
in the chamber. I shall assume that X0 represents a situation in which

the observer sits for one hour and sees nothing. You should note that
since System II omits the radioactive particle, none of the X k 's can
depict the formation of a track due to the event of radioactive decay.
Kutark: Certainly, though, some of the X k 's will represent tracks, for
otherwise Ψ(t)could not represent the observation of any track at all.
Avaroha: That is true. Roughly speaking, for each possible track and each
possible time, there will be an X k that represents the observation of
that track at that time. But in a history X k representing the observation
of a track at time t0, the appearance of the track will be due to some
events other than radioactive decay that take place in System II before
time t 0 .
Kutark: What do you mean? What happens in Xk(t) before t 0 ?
Avaroha: We may not be able to say exactly. Due to the time reversibility
of the Schrodinger equation, we know that X k must have some history
before time t0, and since System II omits the radioactive particle, this
history must depict the formation of the track by some spurious
process.
Referring again to Figure 6, we see that Xn represents a history in
which a complete track appears in the chamber, beginning at time tn.
Each Xk. in the figure represents a history in which a partial track ap
pears at time tki. Starting with Xk1, which represents a very short track,
histories representing successive prolongations of the track at succes
sive times are briefly represented in the total state vector, one after
another. Finally, at time t n the history representing the complete track
becomes represented in Ψ(t).
Kutark: What causes a particular Xk to join the total state vector at a
particular time?
Avaroha: If Xt is already represented in the state vector, and Xj is very
similar to Xt at time t, then at this time it is possible for Xj to acquire a
non-zero Fj and thereby join the total state vector. According to the
mathematics of the Schrodinger equation, 2 4 when X t and Xj are suffi
ciently similar, the wave F, can induce the development of a non-zero
Fj. The situation was similar in the simpler case we discussed earlier, in
which the Xks represented combinations of excited and unexcited
states of target atoms. There we saw that F, could influence Fj only if X,
and A"; differed in the state of excitation of a single target atom.
Kutark: So all the Xks simply exist as unrealized alternative histories, and
when they happen to exhibit certain relationships of similarity to one
another, they can join or leave the total state vector.
Avaroha: That is what the mathematics of the theory indicates. Each of

the histories, Xk1, Xk2, . . . ,Xk involves many events that occur be
fore and after its brief period of "existence" in Ψ(t). These histories
serve only as stepping stones for the establishment of Xn, which rep
resents a complete track. For every possible time and every possible
track direction, there will be a similar system of histories. Thus the to
tal state vector will have a structure we can depict graphically in the
following way. [See Figure 7.] Here each line segment refers to a pos
sible history during its time of representation in Ψ(t). A l l possible
observations of tracks at all possible times are represented. I should
point out that the pattern of branching here is fairly simple because
we are dealing with only one radioactive atom. In a realistic case
with many atoms, each branch would also have many sub-branches
and sub-sub-branches.
Kutark: You are right. The mathematical structure of Ψ(t) is indeed fas
cinating! In the total state vector there are not only "realized" histo
ries depicting different possible experiences of the observer, but there
are also histories that largely involve impossible or highly improba
ble experiences. For example, a history entailing a partial track is
represented only for the short time required for it to serve as a step
ping stone. It is not part of the total state vector once it depicts the
observation of a track that has abruptly stopped!
Avaroha: We might ask, however, just what this mathematical structure
has to do with reality, and how we can correlate its components with
the conscious experiences of the observer.
If we are to avoid contradicting the uniqueness of the observer's con
scious perceptions, we cannot restrict ourselves to performing the op
eration of reduction of the state vector at some one particular time. As
I pointed out, in a case where there are many radioactive atoms, each
branch will repeatedly subdivide, and each time we reduce the state
vector we will have to reduce it again almost immediately.
We are faced with a situation of almost continuously having to make
abrupt, nonphysical modifications in our description of the physical
state of the system, simply to prevent this description from contradict
ing some of the most obvious aspects of our ordinary experience. We
have no basis for deciding at what frequency these modifications are to
be made. We know only that they are necessary if we are to avoid an
absurdity.
Kutark: I can give you an example from ordinary, classical physics in
which such repeated reduction is necessary. Consider a rotating wheel.
The angular position a of the wheel is given by the simple formula
a=vt, where v is the wheel's angular velocity, and we assume that a=0
when the time t is zero.
Now, in physics we cannot define v with perfect accuracy, and due to
the influence of the environment, v will always fluctuate slightly. As
time passes, however, this uncertainty in the definition of v will result
in a large cumulative uncertainty in a. Therefore, it will be necessary to
repeatedly redefine a. Isn't this situation very similar to the one you
are describing? 25
Avaroha: It is different in a very significant way. In your wheel example,
the repeated reduction or redefinition is necessary because our infor
mation about the physical system is incomplete. In the quantum
mechanical situation, this information is supposed to be completely
specified by the state vector, Ψ(t). The very point I've been trying to
make, of course, is that the state vector does not, in fact, provide a
complete description of the system.
X„ X„
x„ x„
Figure 7. The total state vector. Within any small interval of time, many histories
(Xn1"„,, Xn2, Xn3 . . . ) representing different experiences of the observer are added
to^(/). "
I should note that some physicists have tried to make quantum

mechanics complete by introducing what are known as "hidden varia
bles." Their idea is that the state vector corresponds to a statistical dis
tribution over the hidden variables. This means that the state vector is
an incomplete description of the system, and must therefore be peri
odically " r e d u c e d " for the same essential reason that your wheel's
position has to be periodically redefined. As you know, of course,
the hidden variable theories are all very controversial, and none has
met with widespread a c c e p t a n c e . 2 6 T h e y simply constitute another
symptom of the unsatisfactory character of the quantum theory.
Kutark: In practice, however, we can never describe large physical systems
with single state vectors such as Ψ(t). We must always use statistical
distributions, as is d o n e , for example, in statistical mechanics. Also,
we must always consider the effects of minute fluctuations arising from
the influence of the external environment. Thus, in a realistic model
the kind of reduction I have mentioned will always be necessary.
Avaroha: That may well be so. But even if we complicate our model by
replacing Ψ(t) with a statistical ensemble of Ψ j (t)'s. the same basic
problem we have e n c o u n t e r e d with will arise with each of the
individually. This will still be true if we introduce r a n d o m fluctua
tions.27
Kutark: Well, for we do not actually have to perform the operation of
reduction at all. The problems that seem to concern you arise only
when we contemplate the total system, which is described by
When we consider the subsystem lacking the radioactive particle, we
find that one of the Xk(t)'s gives a perfectly acceptable, unambiguous
description of the observer.
Are we not simply confronted here with an example of the profound
and subtle concept of complementarity introduced by Niels Bohr? The
total system and the subsystem are characterized by mutually con
tradictory, complementary features. The total system evolves deter-
ministically and ambiguously in the manner indicated in your Figure7.
In contrast, the subsystem evolves stochastically and specifically, fol
lowing a particular path through this d i a g r a m . 2 8 According to Bohr,
nature is pervaded by such complementary characteristics, and the
realization that this is so is one of the deepest and most far-reaching
philosophical insights t h a t man has yet attained in his quest for
knowledge.
Avaroha: If the total system is ambiguous in certain of its gross features,
then how can a subsystem, which is part of it, be unambiguous in these
very same features?
Kutark: According to the principle of complementarity, when we refer tc

a system we cannot attribute any real existence to its subsystems. This,
in fact, is Bohr's solution to the problem of wave-particle duality. If we
try to attribute real existence to an electron in certain experimental set
tings, then we seem to encounter a contradiction, for it sometimes
exhibits the properties of a wave, and sometimes the properties of a
particle. The contradiction evaporates, however, if we simply refer to
the experimental apparatus plus electron as an indivisible total system.
and do not try to break it down into parts.
Avaroha: Does this mean that when we refer to the total system, we cannot
refer at the same time to the observer, or to his state of consciousness?
Kutark: Yes, that is correct.
Avaroha: But when we speak of System I I , we can refer to the observer as
having specific conscious experiences. Is that right?
Kutark: Yes. You find this confusing only because you do not grasp the
principles of q u a n t u m logic. Just as the general theory of relativity has
forced us to accept new axioms of geometry for our descriptions of
space and time, so the quantum theory has forced us to adopt new rules
of logic. 2 9 It just t a k e s some time to get used to this.
Avaroha: Let me ask you this. Do t h e magnitudes of the F k 's in Ψ(t) have
any bearing on your application of the principle of complementarity to
Let's imagine that by increasing the half-life of the radioactive
a t o m , we increase the relative magnitude of F 0 with respect to all the
other Fk's. Is there some point at which you would d r o p your stricture
of not attributing reality to System II when speaking of System I?
Kutark: As long as the other F k 's are not all z e r o , I can think of no reason
to do this.
Avaroha: In that case, consider what happens when we isolate a single
electron from System II, and designate the remainder of System II as
System III. If we carry out an analysis similar to the one that we have
d o n e for Systems I and II, we will find that the state vector of System II
is a macroscopically ambiguous combination of state vectors for
System III. Do you agree that this is so?
Kutark: It is certainly possible that some unlikely q u a n t u m transitions of
the electron could result in gross variations in the behavior of the
system. Therefore it is reasonable to suppose that the state vector for
System II is macroscopically ambiguous.
Avaroha: Would we also expect to find that the state vectors for System III
are macroscopically ambiguous combinations of state vectors for a
System IV, obtained by removing a particle from System III?
Kutark: Yes. although the relative magnitude of the grossly contradictory
terms in the state vector may be very small. The same thing can be
said of the sequence of subsystems you will no doubt want to construct
by isolating successive particles one at a time. The state vector of any
subsystem may exhibit macroscopic ambiguities.
Avaroha: Doesn't this mean that we cannot associate the unique conscious
experiences of the observer with any subsystem of the total system,
including the man's brain or any part of it? Unless we do perform the
operation of reduction, we are likely to meet ambiguities in any sub-
system that contradict the specificity of conscious experience. It does
not help us very much to simply maintain that whenever we speak
of one of these subsystems, we cannot attribute reality to any of its
sub-subsystems.
Kutark: A l l I can really say here is that we have simply become mired in
useless hair splitting. Since science has to be practical, the best thing to
do is revert to a classical or semi-classical description of the system at
30
some p o i n t .
Yantry: I'm relieved to hear you say that Kutark! This entire discussion of
quantum physics seems very bizarre and unreal to me. In biology and
chemistry we have solid, three-dimensional molecular models that are
as real as your bedroom furniture! We also have practical, compre-
hensible mathematical schemes, such as systems of reaction-diffusion
equations, that describe the behavior of these models. Certainly this
represents the soundest approach to the study of living systems.
Kutark: Yet these mathematical systems can be only approximations at
best. I'm afraid the point will be made that our scientific picture of life
is incomplete if we are forced to resort to such approximations.
3 . 4 A D i s c u s s i o n of Contrasting World V i e w s
(Dr. Sophus Baum and Dr. Francesco Shunya have been listen-
ing for some time, and at this point they join the conversation.)
Baum: I would like to point out that in the midst of all these complicated
deliberations, you have overlooked what is at the same time the
simplest and grandest interpretation of quantum mechanics. In this in-
terpretation we completely reject all ad hoc schemes that attempt to
adjust the state vector in a nonphysical way so as to bring it into line
with our narrow, preconceived views of what reality should be.
Rather, we accept the state vector as a complete description of reality
as it is, with all the staggering implications this entails. Reality consists
of an infinite system of bifurcating, noncommunicating universal
A Discussion of Contrasting World Views 73
realms in which all the manifold possibilities described in the quantum

3
mechanical state vector are factually realized. ' In this interpretation
all transformations of the physical system take place completely in ac-
cordance with the laws of physical causation, as described by the
Schrodinger equation. Chance is completely eliminated in the sense
that a factual universe exists for every alternative state of affairs
described by the quantum theory.
Avaroha: You seem to be referring here to a state vector for the entire
universe.
Baum: Yes, that is correct. Our scientific world view can be complete and
satisfying only if it is universal. Since the world is quantum mechanical,
we must postulate a universal state vector, and the simplest interpreta-
tion of such a state vector is that it describes multiple universes. In fact,
this is the only interpretation of the state vector suitable for the grand
synthesis of general relativity and quantum mechanics that shall
32
provide the final, unified description of universal reality.
Avaroha: Has a mathematically consistent synthesis of quantum mechanics
and general relativity yet been worked out?
Baum: N o , but we are working on it.
Avaroha: Could you explain what happens to the consciousness of the
observer when your universes bifurcate?
Baum: That is very easy to explain. Just as the universe splits into multiple
universes, so the observer splits into multiple observers, one for each
universe. In each universe the corresponding observer is conscious of
the particular course of events that occurs there. In many of these uni-
verses our observer will not even exist. In others, someone similar to
him will exist, and in still others he will exist but will have different ex-
periences. Actually, each one of us has innumerable near-duplicates of
ourselves living separate lives in different universes.
Avaroha: Is it possible for us to communicate with these parallel
universes?
Baum: No. Once a universe bifurcates into two copies, there can be no
communication between them.
Avaroha: Then it is not possible for us to obtain any empirical evidence for
the existence of these universes?
Baum: No. However, this interpretation of the state vector is justified, be-
cause it allows us to retain the absolute statistical laws of quantum
mechanics, which have been verified again and again by experiment.
Avaroha: If every possible quantum mechanical alternative is represented by
a real universe, then there must be universes in which very improbable
sequences of events are the rule.

Baum: I know what you are going to say. The only possible answer is that
we just happen to be in the universe that we are in. Of course, the uni-
verses that obey certain statistical laws are the most prominently repre-
sented, and this is what we mean when we say that these statistical laws
are valid. We just happen to be in a typical universe that obeys the uni-
versal laws, but some of our near-duplicates are now experiencing life
in a universe in which all of these laws have just begun to be completely
violated. That's just the way things are.
Shunya: T h e absurd conclusions you are now contemplating are clearly
due to a serious misunderstanding of the true nature of all scientific
theories. All your attempts to associate some concept of reality with
the algorithms of scientific theories are based on hopelessly muddled
thinking. When making scientific statements, we must speak only
about patterns of correlation in sense data. A theory is of value insofar
as it accurately describes such patterns, and all attempts to associate a
theory with "reality" are simply an irrelevant and meaningless misuse
33
of w o r d s . This is the real import of the quantum theory.
Avaroha: Do your strictures on admissible types of statements apply just to
scientific statements? Could there be true and meaningful statements
of some other kind that do refer to reality?
Shunya: N o . Only statements that refer to patterns of correlation in sense
data should be made.
Avaroha: T h e very statement that you have just made is not of this type.
Your limited system of discourse seems to depend on the existence of a
higher reality for its very definition—a reality to which it cannot refer.
Also, is it possible to live by the system of philosophy that you have
outlined? Do you regard other persons simply as patterns of correla-
tion in sense data, and do you think that it constitutes hopelessly
muddled thinking to suppose that they really exist?
Shunya: The path of scientific truth is a difficult and lonely one. We must
come to the stark realization that most of our desires for meaning, pur-
pose, and understanding can never be satisfied, or even meaningfully
discussed. This applies even to the desire to understand the real basis
of our own existence as scientific investigators. At most we can obtain
some consolation by admiring the beauty of a symmetrical system of
equations, or the elegance of a concise deductive argument. In this way
we can spend the brief interval of time that we somehow pointlessly
occupy between two infinite, meaningless voids.
Kutark: This is becoming very exhausting. I must admit that there seems
to be some inadequacy in our scientific understanding that is prevent-

ing us from coming to a satisfactory conclusion. Possibly this inade-
quacy does have something to do with consciousness. It is true that the
subject of consciousness seems to arise naturally in our discussion, in
various direct and indirect ways. For example, the multiple universe
idea really amounts to a proposal for associating multiple conscious
awarenesses with the state vector. It is also true that we have no idea of
how to deal with consciousness scientifically.
Let me ask you this, Avaroha. No doubt you are aware that some
prominent physicists have proposed that consciousness is directly re-
34
sponsible for the reduction of the state v e c t o r . Some of them main-
35
tain, as you d o , that consciousness is something f u n d a m e n t a l . They
argue that the state vector is somehow reduced at the moment it begins
to display ambiguities that can be perceived by the observer. For ex-
ample, the moment the observer of your cloud chamber begins to
develop an ambiguous brain state representing the perception of two
different particle tracks, the state vector is reduced, and the new state
vector involves just one of these tracks. This is supposedly done by
"consciousness," but how it is done is not at all clear. Some other physi-
cists have even speculated that consciousness can cause systematic
reductions of the wave function that violate the statistical rules of
36
quantum m e c h a n i c s . They use this idea to explain various phe-
nomena such as the "psychokinesis" of the parapsychologists. Are you
perhaps in favor of some of these ideas?
Avaroha: The basic approach that underlies these ideas is to repair the
defects of quantum mechanics by grafting something new onto it. Of
course, it is understandable that a person steeped in the concepts of
modern physics would try to begin in this way. But I would suggest that
this is not the right approach. Someone once pointed out that
Newton's theory of gravitation could not be amended by making some
small alterations, since these would simply clash with the existing
structure of the theory. Rather, Newton's theory could be supplanted
only by something fundamentally different, such as the general theory
of relativity, which provided a completely new perspective. I would
propose that if we are to obtain an adequate understanding of con-
sciousness and its relationship with the physical world, we will similarly
have to approach the subject from a radically different point of view.
Kutark: What do you have in mind?
Avaroha: I will begin by outlining the fundamental understanding of the
nature of consciousness presented in the Bhagavad-gita. As you have
said, even though consciousness is a fundamental aspect of reality,

it has not proven possible to adequately explain or even describe
consciousness within the framework of modern science. The prob
lem, in fact, is not that we presently have an inadequate theory of
consciousness. Rather, we find that our scientific statements do not
even refer to consciousness. This is a fundamental limitation of the
mechanistic point of view, in which we try to reduce all explanations to
mathematical calculations involving measurable quantities.
The Bhagavad-gītā states that each individual living organism
consists of an irreducible conscious entity riding in a physical body
composed of gross material elements. The body is insentient, and it is
described in the Bhagavad-gītā as a complicated machine. 3 7 In con
trast, the conscious entity, or jīvātmā, is the actual sentient self of the
living being, and it cannot be explained in mechanistic terms. 3 8 Each
jīvātmā possesses all of tne attributes of a person, including conscious
ness, intelligence, and innate sensory faculties. These attributes can
not be reduced to the interplay of some underlying impersonal entities
that we might hope to describe by a mechanistic theory. In a sense, we
can compare the jīvātmās with the hypothetical fundamental particles
sought by physicists. Just as these particles are envisioned as having
" certain irreducible material properties, so the jīvātmās can be thought
' of as fundamental units of conscious personality endowed with certain
irreducible personal traits.
Kutark: So, you are proposing a theory that features a transcendental soul
as a primitive element. A priori, I suppose there's no reason why we
can't propose such a thing. But are there any experiments proving the
existence of the soul?
Avaroha: To answer this question, let's consider how physical experi
ments are conducted. In physics we can show that an entity exists only
' by an experiment that takes advantage of that entity's mode of inter
action with other matter. For example, we have been discussing the
• cloud chamber tracks made by charged particles such as electrons.
These tracks result from the ionization of atoms near the path of the
particle, and this ionization comes about as a result of the electro
magnetic interaction between the particle and the electrons in the
atoms. Neutral particles do not interact in this way and thus leave no
tracks. The neutrino, for example, is famous for interacting with mat-
, ter in a very weak fashion—by what is known as the weak force, in
fact—and thus it is very hard to detect experimentally.
Kutark: So to detect the soul in an experiment we would have to take ad-
vantage of the soul's particular mode of interaction. That is reason

able. But what is that mode of interaction? At present, physicists know
of four basic types of force: strong, weak, electromagnetic, and gravi
tational. You are suggesting, I gather, that the soul does not interact in
any of these modes?
Avaroha: That is correct, and consequently we couldn't hope to directly
detect the presence of the jīvātmā with any of our customary physical
instruments. The jīvātmā does interact with matter, but in a very indi
rect and subtle way. As a result of the jīvātmā's primary interaction,
secondary interactions are induced in the body, and these affect the nerv
ous system in particular. 3 9 These electromagnetic effects are detect
able in principle, but they are very complex. Therefore, it would be
very difficult to unambiguously single out the influence of the jīvātmā
on the body by analyzing physical measurements. [The precise nature
of the interaction between the jīvātmā and matter is discussed in
Chapter 7.]
Shunya: This makes it seem very implausible that you could ever verify
the existence of this hypothetical jīvātmā. But I would like to make
another, more fundamental point. Suppose you could demonstrate
from experimental data that some new fundamental law of nature was
needed to explain the functioning of the brain. A theory incorporating
this law would still be mechanistic. A l l our scientific statements—and,
in fact, all valid statements of any kind—refer to patterns in measured
data, and they are therefore necessarily mechanistic, even though we
may not always try to express them in formal mathematical language.
Since you are describing consciousness and personality as non-
mechanistic, you are in effect saying nothing at all. It is meaningless to
even talk about verifying such statements by experiments.
Avaroha: You are partly right. It is indeed true that we cannot study con
sciousness per se by examining its influence on the motion of matter.
Of course, we can make indirect inferences about consciousness by
such methods. But to make a real study of consciousness we have to
take advantage of the higher cognitive potencies of the jīvātmā itself.
As I have pointed out before, many of your own statements cannot
be translated into strictly mechanistic terms. When you spoke of de
sires and hopes that cannot even be meaningfully discussed, you were
in fact referring to aspects of reality that transcend the domain of
mechanistic description. To understand these things and deal with
them in a practical way, we need to enter a domain of discourse and
experience that goes beyond the mechanistic world view. If we were
simply machines, then this would not be possible. However, according

to the Bhagavad-gītā our own existence transcends the mechanistic
realm.
When the jīvātmā is embodied, its innate senses are linked up with
the information processing system of the physical body, and thus the
normal sensory perceptions of the jīvātmā refer almost exclusively to
the physical states of machines, including the machine of the brain. 4 0
In this condition, only the jīvātmā's direct perception of its own inter
nal sensory and cognitive activities involves something that cannot be
described in terms of mechanical configurations. For this reason, there
is a strong tendency for the jīvātmā to overlook its own nature and view
the world in an entirely mechanistic way.
But the inherent senses of the jīvātmā are not limited to observing
the states of the physical body. The jīvātmā is capable of a relation of
„ direct reciprocation both with other jīvātmās and with the paramātmā,
or the Superconsciousness. This mode of interaction directly involves
the use of all personal attributes and qualities, and thus it cannot be
described in mechanistic terms. It can be understood and meaningfully
discussed only by persons who have attained to this level of experience
by direct realization.
Shunya: Such "realization" is purely subjective! Anyone can claim to have
all kinds of remarkable realizations and mystical visions, and in fact
there are many thousands of such people, and thousands of conflicting
sects composed of their gullible followers! But science is limited to
knowledge that can be verified objectively. For an observation to be
considered objective it must be possible for several different people to
(
make the observation independently, and then compare their results.

Avaroha: Two persons able to function on a higher level of consciousness
will certainly be able to recognize each other as realized souls, and they
can also meaningfully discuss their realizations with other similarly en
dowed persons. We can partially illustrate this by the example of two
seeing persons discussing the sunset in the presence of a congenitally
* blind person. The blind person would not be able to appreciate their
statements, and he might take the skeptical viewpoint that talk about
sunsets is simply meaningless. Nonetheless, the seeing persons can dis
cuss the sunset, and each can feel confident that the other is sharing his
experience and understanding what he is talking about.
Another point is that realized persons are able to perceive them
selves and others directly with their innate transcendental senses. They
are not restricted to observations of external behavior. Thus, confir-
A Discussion of Contrasting World Views * 79'
mation of higher states of consciousness is not limited simply to the

subjective perception of each separate person.
You are certainly correct in pointing out that there are many people
who delude themselves and others by claiming to have attained various
kinds of mystical realizations. But the existence of such cheating does
not imply that a genuine science of higher consciousness is not possi
ble. Such a science must indeed be based on verification of cruciat ob
servations by more than one person, but such verification is possible by
realized persons.
The Bhagavad-gītā outlines a practical system for the attainment of
higher realization. In this system the seeker of knowledge must take
instruction from a realized soul. 4 1 By following these instructions, the
person's higher cognitive faculties are awakened by the grace of the
Supreme. 4 2 His realizations, however, can be readily evaluated by his
teacher, who is fully capable of detecting mistakes and illusions.
Furthermore, one can check the conduct of both teacher and disciple
by consulting other self-realized souls, and by referring to a standard
body of authoritative literature. This system can be compared with
that of modern science, in which the findings of individuals are
scrutinized by their peers, and evaluated in the context of standardized
knowledge.
Kutark: You have referred to the Supreme and to Superconsciousness.
What do you mean by these terms? Also, just what is realized in this
process of higher realization? Can this be conveyed to a person whose
experience is limited to the ordinary functioning of the five senses?
Avaroha: The Bhagavad-gītā explains that consciousness exists in two as
pects, the infinite and the infinitesimal. The infinite consciousness is
the very basis of reality and the ultimate source of all phenomena. This
absolute consciousness is understood to be a Supreme Person, who is
fully endowed with all personal attributes—such as senses, will, and
intelligence—and who is known in the Vedic literature by many
names, such as Krsna and Govinda. This is, of course, the same Su
preme Being known as God in the Judeo-Christian tradition, or as
Allah in Islam.
The infinitesimal aspect of consciousness consists of the innumerable
atomic selves called jīvātmās. We can illustrate the relation between
atomic and infinite consciousness by a simple analogy. In classical
physics we can fully characterize an electron as a minute charged parti
cle that interacts in a certain way with an electromagnetic field. Simi
larly, we can understand the true character of the jīvātmā in terms of its
natural interaction with the Supreme Person. Just as we can think of

the electron as being defined by its interaction with an electromagnetic
field, we can understand the jīvātmā as being defined by its personal
interaction with the Supreme Conscious Being.
Thus, the ultimate goal of self-realization is to attain to this state
of natural reciprocation with the Supreme. This mode of interaction
is entirely personal, being based on the exchange of loving service.
We get some hint of the quality of this exchange from the following
characterization of the Supreme Person in the Brahma-samhitā:
premāñjana-cchurita-bhakti-vilocanena
santah sadaiva hrdayesu vilokayanti
yam śyāma-sundaram acintya-guna-svarūpam
govindam adi-purusam tarn aham bhajāmi
"I worship Govinda, the primeval L o r d , Who is Śyāmasundara, Krsna
Himself with inconceivable innumerable attributes, Whom the pure
devotees see in their heart of hearts with the eye of devotion tinged
with the salve of love. " 4 3 Of course, we can attain full understanding of
what this description entails only by direct experience, just as we can
< understand the taste of a fruit only by actually eating it.
Baum: It seems that you are simply expounding a religious doctrine here.
How can you seriously introduce such ideas into a scientific discussion?
From the viewpoint of science, such ideas are simply fantastic and un-
verifiable. You can hold them only by taking a blind leap of faith. To
proselytize these doctrines in the guise of science is reprehensible!
Avaroha: It is ironic that you should say this after advocating the multiple
' splitting universe theory, which is indeed unverifiable. In contrast, di
rect verifiability is the whole point of the approach to higher conscious
ness that I have been outlining. It is certainly true that you cannot
' prove the existence of God on the basis of reasoning and a few pieces of
' empirical data—at least, you cannot prove"anything very interesting
about God in this way. But if God is real and the conscious self is en
dowed with suitable innate capacities, then we should be able to know ,
God directly.
Unfortunately, some scientists have unnecessarily ruled out the very
possibility of doing this by propounding universal theories that leave
no room for the conscious self, what to speak of anything higher. By
extrapolating empirical findings beyond all justifiable bounds, they ar
rive at world views that are truly fantastic and unverifiable. Con
templating such views, they come to the inevitable conclusion that "the
more the universe seems comprehensible, the more it also seems

pointless." 4 4 To widely disseminate such pointless theories among the
people in general is indeed proselytization of a reprehensible kind, for
these theories provide no tangible benefit, and they block the path of
spiritual progress by filling people's minds with misconceptions.
Yantry: This all seems very far removed from the subject matter of phys
ics, chemistry, and biology. Why is it necessary for you to introduce
such an elaborate scheme, which involves so many concepts foreign to
our current scientific way of thinking? Of course, we don't fully under
stand consciousness, and we may have much to learn about the funda
mental principles underlying physical reality. We should certainly be
cautious about the temptation to over-extrapolate and prematurely
generalize. Indeed, it seems to me that as scientists, we should take a
very conservative approach. 4 5 We should simply try to proceed step-
by-step with our research, building gradually on the platform of
existing scientific knowledge.
Avaroha: The trouble with this idea is that modern science doesn't possess
a sufficient body of data about consciousness. A scientific theory is no
better than the data on which it is based, and so progress in under
standing the relation between consciousness and matter is bound to be
very slow and uncertain. This is why the various attempts to introduce
consciousness into the quantum theory are premature. For example,
one physicist may know nothing about consciousness beyond the fact
that persons are normally conscious of definite macroscopic arrange
ments of matter. Since he knows that this fact makes the reduction of
the state vector necessary in quantum mechanics, he proposes a theory
in which consciousness causes this reduction. Yet this theory may fail
to take many aspects of consciousness into account.
The advantage of the Bhagavad-gītā is that it provides extensive in
formation about consciousness, and it also provides scientific proce
dures whereby a person may make practical use of this knowledge.
Since "conservative" means to retain what we already have, a truly
conservative approach to science should not disregard this body of
knowledge.
Kutark: You seem to be trying to build a bridge between a religious
approach to reality and the methodology and world view of modern
science. But aren't these two things completely incompatible? For ex
ample, you are proposing that nature must ultimately be understood in
nonmechanistic terms. If this is true, doesn't it mean that our scientific
approach to understanding nature is futile?
Avaroha: If we adopt a nonmechanistic view of nature, then we must re-

i gard any mechanistic theory as approximate at best. But this will not
put scientists out of business. In principle, the measurable phenomena
generated by nonmechanistic entities can be at least partially described
* by an open-ended hierarchy of mechanistic theories of ever-increasing
46
complexity. Thus there is no lack of available subject matter for
mechanistic theorization. For example, there is no doubt much to be
said in mechanistic terms about the interaction between the jīvātmā
and the physical body, and there is extensive information on this sub
ject in the Vedic literature. In contrast, if it were true that a complete
mechanistic theory could be devised, then once it was worked out in
detail, no further discoveries would remain to be made, and science
would be a closed subject.
Of course, the real interest of the world view presented in the
Bhagavad-gītā lies in the direct study of the jīvātmā and its relation
with the Supreme Person, Śrī Krsna. Since this study is based on re
liable and systematic procedures for obtaining knowledge, it should
certainly appeal to the genuine scientist, even though it completely
transcends the domain of mechanistic theory. This study is of great sig
nificance because it provides a truly satisfying understanding of the na
ture and meaning of the conscious self—an understanding that can be
realized through practical experience.
Notes
1. A. C. B h a k t i v e d a n t a Swami P r a b h u p ā d a , Bhagavad-gītā As It Is.
2. Weisz, Elements of Biology, p. 8.
3. W a t s o n , Molecular Biology of the Gene, pp. 54, 87.
4. S h e r m a n a n d Sherman, Biology: A Human Approach, p. 6.
5. See, for e x a m p l e , Huxley, Essays upon Some Controverted Questions,

p p . 220-221, and Wigner, "Two Kinds of Reality," pp. 248-263.
6. E d w a r d s , e d . , The Encyclopedia of Philosophy, Vol. 5, pp. 336-345.
7. R e n s c h , Evolution Above the Species Level, p. 356.
8. Wigner, " R e m a r k s on the Mind-Body Question," p. 294.
9. Jammer, The Philosophy of Quantum Mechanics.
10. Wigner, "Physics and the Explanation of Life," p. 44.

Notes 83
11. Eigen, "Selforganization of M a t t e r and the Evolution of Biological

M a c r o m o l e c u l e s , " p. 467.
12. Slater, Quantum Theory of Molecules and Solids, Vol. 1, p p . vii-viii.
13. M o t t , " T h e Wave Mechanics of a-Ray T r a c k s , " p p . 79-84.
14. Wigner, " R e m a r k s on the Mind-Body Q u e s t i o n , " p. 286.
15. J a m m e r , p p . 100-101.
16. Wigner, " E p i s t e m o l o g i c a l Perspective on Quantum Theory,"

p p . 379-380.
17. von N e u m a n n , Mathematical Foundations of Quantum Mechanics,

pp. 419-421.
18. H e i s e n b e r g , " T h e R e p r e s e n t a t i o n of N a t u r e in C o n t e m p o r a r y
Physics," p p . 99-100.
19. Rosenfeld, " T h e Measuring Process in Q u a n t u m M e c h a n i c s , " p. 224.
20. von N e u m a n n , c h a p . V.
2 1 . D a n e r i , Loinger, and Prosperi, " Q u a n t u m T h e o r y of M e a s u r e m e n t

and Ergodicity C o n d i t i o n s , " p p . 297-319, and " F u r t h e r R e m a r k s on
t h e Relations B e t w e e n Statistical Mechanics and Q u a n t u m T h e o r y of
M e a s u r e m e n t , " p. 127.
22. Freundlich, " M i n d , M a t t e r , and Physicists," p p . 129-147.
23. Dewitt, " Q u a n t u m Mechanics and Reality," p p . 32.
24. H e r e is a m o r e precise mathematical analysis of t h e state vector. Write

t h e state vector as
w h e r e t h e Xk's represent histories of t h e behavior of t h e a p p a r a t u s

and observer. T a k e {Xk} to be a complete o r t h o n o r m a l set a n d sup
pose, in accordance with o u r assumptions, that each X k satisfies the
Schrodinger e q u a t i o n for t h e system of observer plus a p p a r a t u s (omit
ting t h e radioactive particle). Let F k (x,y,z,t) be t h e wave functions
for this particle for k O. Start at t=0 with Ψ 0 = F 0 X 0 , representing
unexcited target a t o m s a n d no observation of a track. T h e Schr dinger
equation becomes
where V is the potential of interaction between the particle and the

target atoms.
In this equation the operator in brackets on the left represents the
free propagation of the F 's as waves in three-dimensional space. In
k
addition to this free propagation, these waves also directly influence

one another due to the terms (VX ,X ), and thereby transfer weight
k n
from X to histories X with k

0 k 0. T h e wave F can influence the i
wave F only at a time t when X (t ) is nearly identical with X (t ), for

j * t * j *
only at such a time is it possible for (VX (t ),X (t )) to be different

i * j *
from zero. For any given pair of histories this could happen at most
for a brief m o m e n t , and the histories would be expected to diverge
from one another at earlier and later times. Thus weight within the
total state vector will be transferred from one history to another at
times depending on the structure of the histories themselves. This
would result in the development of a total state vector of the form
shown in Figures 4, 6, and 7.
25. Jauch, Are Quanta Real? pp. 4 0 - 4 7 .

26. H e r e is a selection of articles on the theory of hidden variables: Bell,
" O n the Problem of Hidden Variables in Q u a n t u m Mechanics,"
p p . 447-452; Bohm and B u b , "A Proposed Solution of the Measure-
ment Problem in Q u a n t u m Mechanics by a Hidden Variable T h e o r y , "
p p . 4 5 3 - 4 6 8 ; Wigner, "Epistemological Perspective on Q u a n t u m T h e -
o r y , " p p . 377-378; and Wigner, "On Hidden Variables and Q u a n t u m
Mechanical Probabilities," p p . 1005-1009.
27. Freundlich, p p . 132-133.
28. Jauch, p. 89.
29. J a m m e r , chap. 8.
30. H a n s Bethe (personal communication) has pointed out that due to the
problems involved with the reduction of the wave function, it does not
m a k e sense to try to devise a q u a n t u m mechanical description of living
organisms. He recommends an empirical, semiclassical approach.
3 1 . Dewitt, p p . 3 0 - 3 5 .
32. Dewitt, " Q u a n t u m Gravity: the New Synthesis," p. 744.
33. Nagel, The Structure of Science, chap. 6.
34. von N e u m a n n , chap. VI.
Notes 85
35. Wigner, "Two Kinds of Reality."
36. Mattuck and Walker, " T h e Action of Consciousness on M a t t e r : A

Q u a n t u m Mechanical Theory of Psychokinesis," pp. 112-159.
37. A . C . B h a k t i v e d a n t a Swami P r a b h u p ā d a , Bhagavad-gītā As It Is,

p . 830.
38. Ibid., chap. 2.
39. A . C . Bhaktivedanta Swami Prabhupāda, Śrīmad-Bhāgavatam,

3rd C a n t o , Part 4, p. 112.
40. A . C . Bhaktivedanta Swami P r a b h u p ā d a , Bhagavad-gītā As It Is,

pp.702-704.
41. Ibid., p p . 259-260.
42. Ibid., pp. 506-509.
43. Bhakti Siddhanta Saraswati T h a k u r , Shri Brahma-Samhita, p. 124.
44. Weinberg, The First Three Minutes, p. 144.
45. H e i s e n b e r g , Physics and Beyond, p. 112.
46. B o h m , Causality and Chance in Modern Physics, p. 133.

Chapter 4
Karl Popper
On the
Mind-Body Problem
A Review
In The Self and Its Brain, the well-known philosopher Karl Popper and
the eminent neurophysiologist John Eccles have collaborated to develop a
theory of mind-brain interaction. Their book contains three sections. In
the first section Karl Popper does two things: he presents a philosophical
analysis of the mind-body problem, and he outlines a basic conceptual
framework for an understanding of the mind as an entity that is distinct
from the brain but that interacts with it. In the second section John Eccles
surveys our current knowledge of the anatomy and function of the brain
and introduces a specific model for the process of mind-brain interaction.
The third section consists of a series of dialogues between the two authors,
in which they explore some of the strengths and shortcomings of their
ideas.
In this review we discuss Popper's analysis of the mind-body problem. In
his portion of The Self and Its Brain, Popper discusses two basic subject
matters: the nature of mind, and the mind's temporal origin. His aim is to
establish that the fundamental features of personality, such as conscious-
ness, thought, emotion, and purposeful action, are not simply patterns of
chemical interaction (as much current scientific and philosophical thinking
would have it) but rather aspects of a real yet nonphysical entity called the
mind. In Popper's view, the mind should be considered neither an aspect
nor a byproduct of physical processes. Instead, it should be regarded as an
independent entity that influences the physical processes of the brain and
that is influenced by them in turn.
While propounding this bold thesis, Popper simultaneously tries very
hard to show how mind can be understood to have originated in a world of
matter governed solely by physical laws. He does this by elaborating a
theory of emergent evolution, which holds that entirely new and unpre-
dictable qualities, processes, and laws can "emerge" spontaneously in na-
ture and thereby completely transform the character of reality. Although
87
88 Karl Popper on the Mind-body Problem
the idea of emergent evolution is not new, Popper has developed it in great
detail and has made it the cornerstone of his philosophy of mind.
Popper maintains that at one time the universe consisted solely of matter
interacting with itself according to laws similar to or identical with those
studied by the physicists. At this time nothing resembling mind or con-
sciousness existed. Then, as matter gradually coalesced into various combi-
nations by purely physical processes, totally new entities appeared and
began to interact with matter in novel yet lawful ways. Thus life emerged,
then dimly conscious animal minds, and finally fully self-conscious human
minds.
Popper holds that the mind is an emergent entity "utterly different from
anything which, to our knowledge, has previously existed in the w o r l d "
(page 553). He has indeed presented a good case for the view that the mind
is real, that it is independent of the brain, and that its properties are com-
pletely different from those of matter (as understood by the chemists and
physicists). However, in his attempt to reconcile this view with the theory
of evolution, he has shown only that evolutionary theory is unable to ac-
count for the origin of the mind. Although Popper has distorted the stan-
dard neo-Darwinian theory of evolution almost beyond recognition, his
system of emergent evolution contributes nothing to our understanding. In
fact, it raises more questions than it answers.
In the end Popper admits his failure: "Now I want to emphasize how
little is said by saying that the mind is an emergent product of the brain.
It has practically no explanatory value, and it hardly amounts to more than
putting a question mark at a certain place in human evolution" (page 554).
After giving a short account of Popper's arguments for the nonmaterial na-
ture of mind, we will show that emergent evolution is, in fact, logically un-
tenable. Then, salvaging some of the good points of Popper's presentation,
we will consider an alternative approach to understanding the nature of the
mind.
One of Popper's most basic arguments for the nonphysical nature of the
mind is that, contrary to the opinions of the behaviorists, conscious subjec-
tive experience is both real and completely unexplained by our concepts of
matter. The existence of conscious awareness is directly experienced by the
conscious self, and this by itself refutes radical materialism, the view that
matter (as understood by modern physics and chemistry) is all that exists.
Popper also points out that living beings, and especially human beings,
exhibit behavior that cannot be reasonably explained in material terms, but
that can be understood very clearly in terms of the actions of mind. For ex-
ample, it certainly seems very remarkable that a lump of matter in the form
Karl Popper on the Mind-Body Problem 89
of a human being should detach itself from its substrate and, say, climb up
the side of Mount Everest. Although one could dabble with explanations
for this phenomenon in terms of cybernetic brain mechanisms, it is very
hard to see in purely physical terms how or why such mechanisms should
ever have come to be. In general, it is hard to explain long-term and elabo-
rately purposeful action—from the building of nests, to the construction of
skyscrapers, to the formulation of philosophical theories—in terms of
intermolecular forces. Yet we can systematically explain these things in
terms of a mind endowed with purpose and desire.
Popper argues that morality and concepts of good and evil cannot be un-
derstood in purely physical terms. We cannot adequately explain these
manifestations by totally denying the existence of consciousness. We also
cannot explain them by any theory that grants the existence of conscious-
ness, but maintains that it cannot influence matter and that matter behaves
entirely in accordance with known physical laws. (There is a wide range of
such theories, including parallelism, the panpsychism of Spinoza, epiphe-
nomenalism, and the identity theory.) Popper holds that the existence of
standards of moral behavior requires a nonphysical mind that can influence
matter.
The same nonphysical mind is necessary to explain our standards of aes-
thetics. In Popper's words, "To assume that Michelangelo's works are sim-
ply the result of molecular movements and nothing else seems to me very
much more absurd than the assumption of some slight . . . violation of the
first law of thermodynamics" (page 544). He also argues that it is only in
terms of the actions of the mind that we can understand—or even talk
about—truth, falsehood, and validity in language. In purely physical terms
these concepts refer to arbitrary patterns of symbol manipulation, and they
play no essential or meaningful role. They can be meaningfully discussed
only in the context of a reality higher than that of material interactions.
Popper further points out that the mind and its attributes play an explana-
tory role in psychology analogous to the explanatory role electrons play in
physics (although there is, of course, no corresponding mathematical de-
velopment in psychology). Thus it is as reasonable to attribute reality to the
mind as it is to attribute it to electrons. We may add in this connection that
behavioral psychology has provided us with remarkably little insight into
the nature of human behavior, its avowed subject. We might suggest that
psychology will make significant progress only by developing an adequate
understanding of the mind as a real entity.
In refuting the theories based on the idea that consciousness exists but
does not influence the behavior of matter, Popper points out that these
90 Karl Popper on the Mind-Body Problem
theories conflict with the Darwinian theory of evolution, for they do not
provide a means whereby natural selection can act to bring about the evolu-
tion of consciousness. Thus the strict Darwinist must either deny the reality
of consciousness or attribute full consciousness to inanimate matter. Other-
wise he must explain why consciousness should develop progressively with-
out the aid of natural selection. This argument should be of interest to the
orthodox adherents of neo-Darwinian evolutionary theory. As we shall
see, Popper himself proposes a very different theory of evolution.
Popper also confronts the challenge that since the known laws of physics
cannot be violated, a nonphysical mind must be ruled out. His main point
here is that the program of reductionism in physics and chemistry is drasti-
cally incomplete. Because of difficulties in computation, we simply do not
know what the quantum theory says about most chemical processes. Also,
there have been many revolutionary developments in physics in this
century, and we may confidently expect more.
Adding a final point, Popper warns of the danger of cutting off healthy
inquiry by prematurely imposing a rigid view based on known physical
principles. In philosophy, this warning applies to the indiscriminate appli-
cation of'Occam's razor as a means of ruling out hypotheses. Popper also
condemns what he calls "promissory materialism," or the argument that
materialistic claims should be accepted because proof for them will surely
be found some time in the future.
Thus Popper gives many arguments indicating that mind must involve a
higher reality—a reality that cannot be reduced to matter as it is under-
stood by modern science. However, when he tries to account for the origin
of this higher reality by means of his theory of emergent evolution, he sticks
essentially to the reductionistic approach, and he is unable to avoid its
shortcomings. We will now briefly outline the theory of emergent evolution
and indicate why this is so.
First we shall consider Popper's concept of "downward causation."
According to this idea, when many parts combine to make a whole, the
whole is an entity in and of itself, and it is capable of acting on each of its
parts individually. The formation of heavy nuclei in the center of a star by
the action of intense heat and pressure is regarded as an example of down-
ward causation, in which the whole star acts on its individual constituent
particles.
This concept amounts to a deceptive way of introducing higher entities
into the material picture without really introducing them. A physicist
would immediately reply that the behavior of the star can be fully explained
(in principle) in terms of the material interaction of its component parts,
without any reference at all to the star as a whole. Thus the concept of
downward causation is simply a disguised form of reductionism.

In fact, Popper does believe that new entities can appear in the world,
and that these entities amount to more than simply names for collections
of existing entities. In particular, he believes entirely new laws or modes
of operation of the universe can come into being, and thus "the first
emergence of a novelty such as life may change the possibilities or propen-
sities in the universe" (page 30). Popper maintains that the inherent inde-
terminacy of quantum mechanics makes it possible for new laws to appear
without violating the old ones. Thus he believes that "all physical and
chemical laws are binding for living things" (page 36), even though living
things involve the "emergence of new lawlike properties" (page 25).
Popper does not like the idea of substance or essence, and thus he tends
to refer to his emergent entities in terms of their behavior. However, in ad-
dition to exhibiting novel patterns of lawful behavior, emergent entities
like life and mind possess, according to Popper, entirely new qualities and
properties—such as consciousness. Thus they add to nature something
concretely new, and very much like a substance or essence.
Although Popper believes the appearance of these entities to be com-
pletely unpredictable, even in principle, he still claims that their appear-
ance is lawful. Accordingly, Popper repeatedly stresses natural selection as
a basic principle underlying the development of mind, although mind is a
totally novel, emergent entity that cannot be reduced to a combination of
material elements. At the same time, Popper agrees with Jacques Monod
when he says that the chance that life would appear on earth was virtually
zero, and he cites this as an example of how emergent entities are com-
pletely unpredictable in terms of the universe as it exists before the time of
their emergence.
Generally, Popper gives the impression that emergent entities appear
gradually in a step-by-step fashion over the course of evolutionary develop-
ment. According to this view, mind as a fundamental entity in its own right
must have emerged at a certain stage of evolution, thereby altering the
basic character of nature from then on. However, in their dialogues Popper
and Eccles also seem to agree that mind emerges anew from matter in the
brain of each newborn child—a position that Eccles supports by saying that
a girl who had been brought up for 1 3 ½ years without learning how to
speak did not have a mind during that time (page 564). This appears to
be contradictory, but in any event Popper seems to require that the
emergence of mind must be lawful.
We can raise many objections to this theory of emergent evolution, and it
is clear that Popper has not formulated it consistently. However, we will re-
strict ourselves to making two basic points. The first is that the theory of
Karl Popper on the Mind-Body Problem
emergent evolution is very different from the standard neo-Darwinian

theory of evolution, and thus it should be understood that Popper is actu-
ally rejecting the standard evolutionary theory. The second point is that for
new entities to emerge in a lawful fashion, as Popper certainly requires,
, there must be a preexisting higher law that regulates their emergence. This
higher law must already incorporate, in potential f o r m , the qualities and
features of the emergent entities.
As a concrete example, suppose that the state of a physical system is
1
changing according to a Markovian probability law, L. The state X of such
a system makes random jumps at successive discrete times, t= 1, 2, 3 , . . . ,
and these jumps are governed by the law L. The behavior of the system is
summed up by the rule that at each time t,
X > X' with probability L(X,X')
For example, suppose that A'represents the position of a drunkard who is

staggering back and forth at random on a steep hillside. Then the formulas,
L(X,X - 1)=2/3 and L(X,X +1)=1/3, would mean that the drunk always
takes a step of one unit, but is twice as likely to step downhill as uphill.
Now let us imagine that the system can develop an emergent property
characterized by a new law, L ' , and let us suppose that in the state X, this
emergence can take place with a probability of p(X). If we try to formulate
this, we can see that it entails a metasystem in which both L and X are vary-
ing. The behavior of this enlarged system, if it can be described at all, must
be given by a more complex meta-law. For example, we might stipulate
that at each time t,
(X,L)—>(X',L') with probability L(X,X') L(X,L,L')
This meta-law necessarily entails a function L(X,L,L') giving the proba-

bility that the law V will emerge when the system is in the state X and is
operating under the law L. In our example, all of the possible emergent
2
laws L' can be expressed by formulas, and thus L can also be represented
mathematically. But what can we say about the meta-law L in Popper's
system, where mental properties may emerge that are qualitatively differ-
ent from anything contemplated in present-day physics? Certainly this
meta-law would also have to be qualitatively different from any of the laws
known to physical science. Indeed, if anything, it would have to possess
qualities even more remarkable than those of the various possible entities
that may emerge under its regulation.
Now, if the meta-law L is not itself an emergent entity, it must represent
a permanent state of affairs, and its qualities must correspond to absolute
features of reality. If L is an emergent entity, we can inquire about the law

governing its emergence, the law governing the emergence of this law, and
so on. Unless we propose an infinite regress of laws, we must posit some ul-
timate, absolute law that can be discussed only in terms of nonphysical
entities and qualities.
Thus Popper's theory of emergent evolution requires the existence of
laws that already embody qualities equal or superior to all the supposedly
novel qualities and properties that may "emerge." Such laws must be in-
voked to account for lawlike regularities in the emergence of new qualities,
and to explain, for example, why we should expect mind to emerge from a
highly organized brain and not from a clod of earth.
This conclusion certainly contradicts the basic purpose of Popper's
system of emergent evolution. We would like to suggest that this system
should, in fact, be discarded, for it is an attempt to carry out the impossible
task of accepting reductionism and rejecting it at the same time. A more
satisfactory development of Popper's ideas about mind can be obtained if
we boldly reject reductionism altogether and simply accept that principles
and entities higher than those known to modern science are, and always
have been, a permanent part of the world.
Two aspects of our world that are particularly significant are the pres-
ence of conscious selves, and the presence of standards of meaning and
value. These are certainly of central importance, for it is ultimately the self
that understands and seeks to find explanations, and in so doing the self is
always referring to standards of meaning. It is therefore interesting to note
that Popper comes close to attributing real, irreducible existence to the self
and to absolute standards of meaning.
Thus Popper describes something he calls "World 3," the realm of ab-
stract concepts, mathematical truths, and moral and ethical principles.
"World 3 is a kind of Platonic world of ideas, a world which exists no-
where but which does have an existence and which does interact, especially
with human minds" (page 450). Popper stresses the role of World 3 as a
final standard of reference for questions of truth, validity, and meaning
(page 77).
If there is no fundamental basis for meaning, then it is hard to see how we
can ultimately justify any philosophy of life, on either a theoretical or prac-
tical level. However, if we grant the existence of a factual, absolute source
of standards for truth and meaning, then a possibility opens up—we may
be able to gain access to this source and thereby obtain a genuine under-
standing of the world. A n d if we at least tentatively accept the self as a real
entity, then we can entertain the possibility that the self may have a mean-
ingful place in the larger order of reality as a whole. A direct investigation
94 Karl Popper on the Mind-Body Problem
of the role of the self in the context of a higher, meaningful reality could
lead to practical realization of our true potentialities as conscious beings—
potentialities that presently lie dormant and unexplored.
Popper tends to speak of the self in rather vague reductionistic terms as
some kind of pattern of interaction between the mind and the body. Yet he
also sometimes refers to the self as a real entity. In one passage he says that
"the talk about a substantial self is not a bad metaphor" (page 146), and he
proposes that "selves are the only active agents in the universe: the only
agents to whom the term activity can be applied" (page 538). This is really
Popper's best bet, and one wishes he had the boldness to follow his idea
where it leads and straightforwardly consider the self as a fundamental
entity in its own right.
On the basis of such a hypothesis, a direct investigation of the self as such
is at least a formal possibility. We suggest that only an investigation of this
kind can yield satisfying answers to the basic questions that gave rise to The
Self and Its Brain in the first place. As Eccles points out in one of the
dialogues, " M a n has lost his way these days . . . He needs some new mes-
sage whereby he may live with hope and meaning. I think that science has
gone too far in breaking down man's belief in his spiritual greatness and in
giving him the idea that he is merely an insignificant material being in the
frigid cosmic immensity" (page 558). If we can candidly consider that the
self may be a real, conscious entity in a world with inherent meaning and
purpose, we may hope to come to a practical and coherent world view that
does violence neither to reason nor to our natural spiritual aspirations.
Notes
1. A simple Markov chain is used in this example, but the same argument
could be framed using the more sophisticated laws studied in physics.
2. There are many different ways of formulating such a meta-law. The
basic point is that the meta-law must govern the emergence of certain
"novel" properties, and thus it must be able to deal with these proper-
ties. A law defined in the customary language of physics could not gov-
ern the emergence of consciousness, since consciousness belongs to a
category lying outside the scope of this language. If we suppose that
there are laws of nature governing the emergence of consciousness,
then we must conclude that nature has inherent features falling outside
the realm of discourse of physics. The simplest hypothesis about such
features is, of course, that consciousness itself is an inherent feature of
reality, and not an emergent property.
Chapter 5
Information Theory
And the
Self - organization
Of Matter
In this chapter we use information theory to show that the laws of nature,
as understood by modern science, are insufficient to account for the ori-
gin of life. The basic argument is the following: The laws of nature and
the corresponding mathematical models of physical reality can all be de-
scribed by a few simple equations and other mathematical expressions.
This means that they possess a low information content. In contrast, there
is good reason to suppose that the intricate and variegated forms of living
organisms possess a high information content. It can be shown that con-
figurations of high information content cannot arise with substantial proba-
bilities in models defined by mathematical expressions of low information
content. It follows that life could not arise by the action of natural laws of
the kind considered in modern science.
This argument bypasses the proposition made by many evolutionary
theorists that even though the steps leading to life are improbable, they are
nonetheless likely to happen, given the immense spans of geological time
available. We show that no period of time, from zero to billions of billions
of years, will suffice to render probable the evolution of life from matter by
known physical processes. Indeed, we show that over periods many times
the estimated 4.5-billion-year age of the earth, the probability of the evolu-
- 1 5 0 , 0 0 0
tion of higher life forms remains bounded by upper limits of 1 0 , an al-
most infinitesimal number. This implies that the entire history of the earth
1 5 0 , 0 0 0
would have to be repeated over and over again at least 1 0 times for
there to be a substantial chance that higher living entities would evolve
even once.
At the basis of these figures lies an intuitive reason for the impossibility
of organic evolution. We show that the process of natural selection—the al-
leged mechanism of evolution—must have specific direction if it is to bring
about the development of complex living organisms. Without such direc-
tion, this process is unable to discriminate among random events
(mutations) in such a way as to bring complex order out of chaos. For this
98 Information Theory and the Self-organization of Matter
reason, the standard argument that evolution will occur, given long enough
time spans, is false. Natural selection will lack direction if the causal prin-
ciples underlying this selection, namely the laws of nature, do not have
sufficient information content to specify such direction.
Our fundamental argument is that in a physical system governed by sim-
ple laws, any information present in the system after transformations cor-
responding to the passage of time must have been built into the system in
the first place. Random events cannot give rise to definite information,
even when processed over long periods of time according to simple laws. It
follows that mathematical models based on simple laws cannot account for
the origin of such highly variegated and complex entities as living organ-
isms. Using such models, we can only explain the existence of complex
order here and now by postulating that equivalent complex order was pres-
ent at an earlier time, or was transported into the system from the outside.
These postulates do not account for the origin of such order, but simply
confront us with either an infinite regress or an eternal source of order
containing the information necessary to specify all life forms.
These arguments suggest that the scientific program of describing the
world by mathematical models is severely limited, and can serve only to im-
pede our understanding of nature. This program is based on the conviction
that the simple regularities observed by physicists and chemists in experi-
ments with inanimate matter will suffice to account for every phenomenon
in the world. But since we show here that this program cannot succeed in
accounting for the origin of life, we suggest that a different approach
should be explored. The direct implication of the information theoretic
analysis is that nature is inherently complex, and encodes the designs of liv-
ing organisms—both lower and higher—in some form. Although this in it-
self does not tell us very much, we suggest that it may be fruitful to consider
the possibility that life is built into nature as a fundamental principle, and
represents more than simply a temporary collocation of material elements.
The program of science during the last two or three hundred years has
been to reduce life to matter and to deny the existence of any higher life-
principle transcendental to matter. In this program, the idea that matter is
both simple and conceivable has been essential. But if, as shown here, we
must attribute all the characteristic features of life to matter in order to ex-
plain the origin of life from matter, then we can conclude that this program
has failed.
One way of responding to this failure is to consider the universe as a
whole to be an incomprehensible welter of arbitrary complexities. A more
promising response is to consider that if the natural causes underlying life
must entail all the detailed features of life, then perhaps life—and intel-
The Theme of Simplicity in the Theories of Physics 99
ligent life in particular—is the fundamental cause underlying the phe-

nomena of the universe. This possibility opens up many new avenues of
empirical investigation, but it also implies that traditional empirical
methods can take us only so far in our effort to understand nature. Yet, if
the ultimate cause of the universe is not merely " l i f e , " but is an irreducible
intelligent being, then the possibility arises that we may be able to tran-
scend the process of experimental dissection and hypothesis, and obtain
knowledge by direct communication with this absolute source.
In this chapter we touch on these points only briefly and concentrate on
the analysis of natural processes using information theory. We develop the
basic arguments in Sections 5 . 1 , 5.2, and 5.3, and outline their broader
implications in Section 5.4. A more detailed mathematical treatment of
some of the arguments is presented in Appendices 1 and 2.
5.1 The Theme of Simplicity in the Theories of Physics

One of the fundamental principles of modern physics has been that the
laws of nature can be described by very simple and general mathematical
relationships. Perhaps the most famous advocate of this principle was
Albert Einstein, who strove during the major portion of his life to find a
"unified field theory," which would derive all the forces and laws of inter-
action of the physics of his day from a single, unifying mathematical rule. In
more recent times, the search for a unified physical theory has been avidly
pursued by many physicists, such as Steven Weinberg, Abdus Salam, and
Sheldon Glashow, who in 1979 shared the Nobel Prize in physics for their
contributions in this field.
Weinberg and his colleagues have not yet been able to formulate the
"grand unified theory" with which they hope to pin down, once and for
1
all, the ultimate reality underlying the phenomenal w o r l d . Such a theory
would account for the observed behavior of subatomic particles at high
energies and would provide a complete theoretical description of the ori-
gin of the universe from a primordial cosmic explosion, or "big bang."
Although many researchers have tried to construct such a theory, serious
mathematical difficulties have thus far impeded their efforts. For exam-
ple, the quantum field theory is plagued by a tendency for important cal-
culations to diverge to infinity when they should yield finite answers. Also,
no one has yet been able to devise a consistent mathematical system that
joins together quantum mechanics and general relativity, the two most
2
fundamental theories of modern physics.
As a result of these difficulties, it is not possible at the present time to
formulate a complete mathematical model of the universe. But scientists
widely believe that a complete physical theory has been established for
phenomena involving moderate masses, energies, and velocities. In par-
ticular, scientists generally believe that the nonrelativistic theory of quan-
tum mechanics is sufficient to account for all the phenomena of chemistry,
and that all the phenomena of life can be reduced to chemistry. (Nonethe-
less, some long-standing controversies about the validity of quantum
mechanics remain unresolved. These were discussed in Chapter 3 . )
According to the physicist John Slater, "The success of exact calculations
for the helium atom and the hydrogen molecule convinced physicists that
wave mechanics provided a framework which was at least in principle capa-
ble of giving theoretical explanations of any desired accuracy for the phe-
3
nomena of atoms, molecules and solids." The idea that life, in turn, is a
molecular phenomenon is summed up by the molecular biologist James
Watson: "We see not only that the laws of chemistry are sufficient for un-
derstanding protein structure, but also that they are consistent with all
known hereditary phenomena. Complete certainty now exists among
essentially all biochemists that the other characteristics of living organisms
(for example, . . . the hearing and memory processes) will all be com-
pletely understood in terms of the coordinative interactions of small and
4
large molecules." Taken together, these statements of Slater and Watson
imply the widely accepted conclusion that life is a quantum mechanical
phenomenon.
The laws of quantum mechanics must be very remarkable if all the char-
acteristics of living beings do indeed depend on them. We will therefore
briefly outline the mathematical structure of the quantum mechanical laws.
Our objective is twofold: first, to clearly define what we mean by simplicity
in a system of natural laws; and second, to clarify the nature of the
hypothetical relationship between the quantum mechanical laws and the
phenomena of life.
First, let us consider the laws of nature in classical physics. These can be
summed up by the following equations:
In classical physics the state of a physical system at any given time is

completely described by the position coordinates, and momentum coor-
dinates, p . These coordinates comprise 6N numbers for a system of N ma-
j
terial particles, and equations (1) and (2) describe how these numbers
change with the passage of time. In classical physics the function H, which
is called the Hamiltonian, is generally given by a simple formula:
We have written these formulas out in full to show how very simple the laws
of classical physics are. This is quite literally the full extent of the laws of na-
ture as understood in classical physics up to the time of Maxwell. Accord-
ing to those scientists who adhered to the philosophy that nature could be
completely described by mathematical laws, all the phenomena of nature
could be calculated using equations (1) through (5) and initial values of the
q 's and p 's at some arbitrary starting time, t = 0 .
j j
This philosophy received its initial impetus in the eighteenth century

from Isaac Newton. He summed it up as follows: "I . . . suspect that [the
phenomena of nature] may all depend upon certain forces by which the
particles of bodies . . . are either mutually impelled towards one another
and cohere in regular figures, or are repelled and recede from one
5
another." Equation (4) specifies these forces, which attract if A is posi-
ij
tive and repel if it is negative. In the nineteenth century the physicist

Hermann Von Helmoltz expressed the same view: "The task of physical
science is to reduce all phenomena of nature to forces of attraction and
repulsion, the intensity of which is dependent only upon the mutual dis-
tance of material bodies. Only if this problem is solved are we sure that na-
6
ture is conceivable." Needless to say, Von Helmoltz included life as a
"phenomenon of nature."
With the advent of Maxwell's electromagnetic theory, Einstein's theory
of relativity, and the theory of quantum mechanics, the simple view of na-
ture summed up by equations (1) through (5) underwent a considerable
change. But the basic concept—that natural phenomena are reducible to

the interplay of elementary material forces—has been retained. In the
present dominant theory of quantum mechanics, arrangements of numbers
still describe physical systems, although the particle coordinates of classical
physics have given way to Hilbert space vectors. The laws of transforma-
tion of these numbers are still given by simple equations that may be
written down in a few lines.
In quantum mechanics, the basic equation of motion for a physical
system is the Schr dinger equation:
Here the state, or exact physical description, of the physical system is given
by the Hilbert space vector, which can be represented in various ways as
a mathematical function or as a sequence of numbers. The Hamiltonian
function, H, has been adopted from classical physics and now appears as an
operator capable of acting on to produce a new vector. In analogy with
equation (3), H could be given by:
where V is the same as in equation (4).

Equations (6) and (7), along with (4) and (5) and an initial value for W
at the time t=0, completely specify the quantum mechanical picture of
a physical system of n/3 particles moving according to the attractive and
repulsive forces given by V.
In further developments of the theory of quantum mechanics, things
become somewhat more complicated. In addition to equation (4), various
other terms are added to V to represent different kinds of forces believed to
be acting in physical systems. These include terms for "spin" and electro-
magnetic interactions. Also, the basic form of H in equation (7) is some-
times modified in various ways. It remains true, however, that very brief
formulas can express the Hamiltonian for any system that is supposed to
represent the fundamental laws of nature. When the abbreviated notations
in these formulas are written out in full (as we did for the classical case) the
resulting equations occupy a few lines at most.
This is true in particular of the physical model of chemical interactions
that Watson and his colleagues believe to be sufficient for a complete un-
derstanding of life. The Hamiltonian for this model should include terms
for electric forces, spin interactions, and electromagnetic interactions (plus

7
gravity). This Hamiltonian is illustrated in Figure 1.
If the laws of physics can account fully for the moment-to-moment flow
of life processes, one might suppose that they also can account for the ori-
gin of living organisms over the course of time. In the development of mod-
ern science, Charles Darwin was one of the first persons to seriously work
out the consequences of this idea. Darwin based his theory of evolution on
hypothetical processes of random variation and natural selection that he
regarded as natural byproducts of underlying physical processes. Since
Darwin's time, biologists have ascribed his random variations to the pro-
cesses of genetic mutation and recombination. They have also carried his
ideas to their logical conclusion by introducing the theory of molecular
evolution. This theory attributes the origin of the first living organisms to
purely physical processes of "self-organization" acting in a primordial soup
of disorganized chemical compounds.
It is our thesis that equations as brief and simple in form as those of Fig-
ure 1 cannot contain sufficient discriminatory power to summon forth from
a chaos of randomly distributed atomic particles the complex and varie-
gated world of life we see about us. Both the Darwinian and the molecular
Figure 1. The laws of nature underlying chemistry.

theories of evolution are based on the processes of random variation and

natural selection. These theories hold that "chance" will provide various
combinations of molecules which may or may not be useful to living or-
ganisms, and that "natural selection" will pick out those which are useful
and eliminate those which are not. Geneticists such as Ronald Fisher have
argued statistically that even if natural selection only slightly favors one
form over another, in a sufficient length of time the favored form will
8
nonetheless be found to predominate over the unfavored one.
Yet natural selection must have some direction if it is consistently to
choose certain material configurations out of the myriads of configurations
possible. The local selective advantages within particular populations must
add together (as in a vector sum of vectors added tail to head) to produce a
general trend from primordial soup to higher organisms. Ultimately, the
fundamental laws of nature must provide this direction. At least, this must
be true if nature is indeed to run in accordance with such laws.
It is very hard to see, however, why "forces of attraction and repul-
sion . . . dependent only upon the mutual distance of material bodies"
should select trees, amoebas, bumblebees, or human beings in favor of
other possible material configurations, such as inert globs or blotches En-
hancing the theoretical picture with spin interactions following the Pauli
matrices, or electromagnetic fields composed of harmonic oscillators^does
not seem to add more plausibility to the idea that natural selection could do
this. We will argue that, in current physical theories, the very brevity of the
laws of nature makes them incapable of selecting the complex forms of liv-
ing organisms from an initial state of molecular chaos, no matter how much
time is allowed for the process. Basically, we shall show that for natural
laws to select a complex form out of a random distribution of matter, the
laws must themselves possess a corresponding level of complexity. This
implies that the Hamiltonian for a quantum mechanical system will require
many pages of symbols (as many as thirty at the very least) if that system
is to be capable of selecting configurations with the complexity of living
organisms. In other words, the known laws of physics are insufficient to
account for the origin of life, and for a system of physical laws to do so,
its sheer size and complexity would make it impossible for the human mind
to handle.
In neither the classical nor the quantum mechanical theories of physics
have mathematicians ever been able to exactly solve the equations of mo-
tion for systems containing more than two particles. Since systems capable
of representing living organisms must contain enormous numbers of parti-
23
cles (on the order of 1 0 ) , we can see that it is not practical to study the
nature of such organisms by explicitly solving the equations of motion.

But such solutions must exist for the theory to be valid at all. Generally
scientists have argued on the basis of abstract reasoning that solutions to
the equations of motion can be calculated in principle. They have then
attempted to apply and verify the theory by making logical deductions
about the properties of the solutions without actually seeing them. We shall
proceed on the assumption that the required solutions always do exist and
could be calculated.
We shall formally define the information content of a theory to be the
length of the shortest computer program that can numerically solve the
equations of motion for the theory to within any desired degree of accu-
racy. For consistency, we shall stipulate that all programs used in the esti-
mation of information content must be written in a fixed programming
language. The Schr dinger equation (Equation 6) can be solved in prin-
ciple by a simple numerical algorithm. Consequently, the information
content of a theory having this equation as its basic equation of motion
is nearly proportional to the number of symbols needed to write out the
Hamiltonian for that theory in the programming language. We can also es-
timate the information content of a configuration of matter, such as the
body of a living organism, to be the length of the shortest program needed
to generate a complete numerical description of that configuration. We
shall use this measure of information content to provide a clear-cut quanti-
tative demonstration that the known laws of physics, or any similar system
of laws, must fail, even in principle, to account for the origin of life.
In a mathematical model of a physical system, two other ingredients are
needed along with the laws of motion. These are the initial conditions and
the boundary conditions of the system. Normally, a physical system is con-
fined within a certain fixed volume of space. To calculate the events occur-
ring within this volume, we must know the physical conditions along the
boundary of the volume during the time the system is being studied. In
most physical models the behavior of the system is studied during a finite
time interval, 0 t t . It is therefore also necessary to specify the physical
1
state of the system at the beginning of this time period, or t=0.

In a theory of the origin of life, the initial conditions typically represent a
"primordial" situation possessing a very low degree of organization, if any.
For example, most theories of the chemical origin of life postulate that life
arose from a "primordial soup" consisting of a mixture of water and simple
compounds such as C 0 , C H , N , N H , and H S , and a reducing atmos-
2 4 2 3 2
9
phere composed mainly of C H and N H . This mixture of chemicals is
4 3
presumed to receive radiation from the sun, to receive supplies of gases

from the earth (volcanic venting), and to radiate heat and light into outer
space. This sums up the initial and boundary conditions for this model.
As another model, one could start with the supposed origin of the solar
system from a cloud of gas. The initial conditions would then consist of a
description of the initial gas cloud, and the boundary would correspond to
an unlimited vacuum surrounding this cloud (if we ignore the influence of
distant stars). In this model the laws of nature would first generate the solar
system, complete with primordial soup, and then generate life in the soup.
Or one might consider a model of the universe as a whole, such as that pro-
posed by the "big bang" theory. Most versions of this theory feature a
superhot soup of subatomic particles as the initial condition of the uni-
verse. (Of course, there are no boundary conditions in a universal model.)
We should observe that the quantum mechanical laws of Figure 1 are not
adequate for the last two of these three models. Since the second model in-
cludes the sun, it must deal with nuclear reactions, and these are not de-
scribed by nonrelativistic quantum mechanics. This model would require
some form of quantum field theory. Likewise, the third model calls for a
theory combining quantum mechanics and general relativity. It will not be
possible to precisely define these models until substantial advances are
made in theoretical physics. Yet it is interesting to note that the existing
equations of quantum field theory are much more streamlined and elegant
than the nonrelativistic equations of Figure 1. Since the general program of
physics is to explain nature in terms of the simplest possible principles, we
can anticipate that physicists will exert every effort to devise universal
equations of motion that are even simpler in form than the quantum field
equations.
We should also observe that the initial and boundary conditions of the
three models are quite simple to describe and that they become progres-
sively simpler as we go from model to model. The basic idea behind current
scientific theories of the origin of life from matter is that one need propose
only a very simple set of conditions to hold in the beginning. After all, the
purpose of such theories is to "explain" all the features of life, and the more
intricate the specifications required for the initial conditions, the less com-
plete the explanation becomes. Many scientists would feel a need to ex-
plain complex initial conditions in terms of some earlier and less complex
state of affairs.
In this regard it is interesting to consider in greater detail the initial con-
ditions that scientists have proposed for the big bang theory. One of the
basic requirements of this theory is that it should account for the distri-
bution of matter in the universe into galaxies and clusters of galaxies. Yet
scientists have noted that an initial superhot plasma will tend not to give
rise to such a distribution. At high temperatures, diffusion will tend to

smooth out all irregularities in a gas, and the later development of a highly
irregular distribution of matter can be assured only by building very large
irregularities into the initial plasma. Some theoreticians, including the
astrophysicist David Layzer, have felt dissatisfied with this idea. They have
introduced an alternative theory in which the universe begins in a perfectly
homogenous state of global thermodynamic equilibrium at zero tempera-
10
t u r e . In their theory, the irregularities in mass distribution are produced
by the fracturing of the solid universal substance as the universe expands.
Their initial conditions are thus as simple as possible, and they require
that all diversity in the universe must arise from the action of processes
governed by the natural laws.
A typical initial condition for a model of the origin of life will consist of
an ensemble of possible initial states, such as one of the thermodynamic en-
sembles of statistical mechanics or some simple combination of these. This
certainly holds true for the models referred to above. Such an ensemble can
be specified by a simple equation or a brief set of equations. For example,
in the theory of quantum mechanics one of the standard thermodynamic
ensembles, known as the canonical ensemble, is given by the equation
po = K -l
exp(-H/kT) (8)
11
where H is a Hamiltonian operator.
In this equation, p is called a density matrix. It corresponds to a col-
0
lection, or ensemble, of quantum mechanical states, each of which has a

statistical weight giving the probability that the physical system will be
found in that particular state. This means that the initial state of the system
is left ambiguous to a very large extent. In fact, one of the basic principles
of statistical mechanics is the assignment of equal probabilities to all ini-
tial states satisfying certain simple restrictions, such as a certain particle
density or a certain range of energies. In such statistical formulations of
initial conditions, many of the alternative initial states of the system will
have a very high information content. However, the statistical ensemble as
a whole can be described by a few simple criteria, and therefore it possesses
a low information content.
The boundary conditions for the physical system should also be defin-
able in simple terms. At most, one would expect radiant energy or simple
material particles to pass back and forth across the boundary. In some
models the boundary might consist of reflecting or noninteractive walls, or
a limitless vacuum, or there may be no need for explicit boundary condi-
tions at all. Any interaction between the system and matter or energy
beyond its boundary should be describable in simple statistical terms, as
can be done, for example, for cosmic rays or the influx of solar radiation.
Boundary conditions, like initial conditions, must be relatively simple: if
intricate specifications were required for the boundary conditions, then
one would need to explain their origin also.
Once the initial and boundary conditions are defined, the state of the
physical system at each time t between 0 and t can be calculated using the
1
natural laws of the system. The basic situation is summed up in Figure 2.

A theory of the origin of life from matter should predict that after some
period of time, perhaps in the range of four billion years or so, the sys-
tem will have a reasonably high probability of exhibiting the molecular
configurations characteristic of living organisms.
Let us consider how we can mathematically represent the presence of
these configurations in the physical system. We will do this in particular for
the theory of quantum mechanics. In this theory, any state of affairs one
might want to observe in the system is represented by a mathematical
operator called an "observable." We can denote the situation to be ob-
served by B,, and we can denote the mathematical operator to which it cor-
responds by B. Then the probability of finding the system in the situation B
is given by
Prob(B) = Trace(pB) (9)

where p is the density matrix defining the state of the system. This formula
can be evaluated in principle by a brief set of computer instructions.
We are interested in the case where B represents the presence of a par-
ticular molecule or collection of molecules in the system. For example, B
might represent the presence of a D N A molecule expressing the genetic
specifications of an organism. One way of describing such a molecule is by
means of a numerical code indicating which atoms are to be found in close
proximity to which other atoms—that is, indicating the pattern of chemical
bonds that characterizes the molecule. If X represents such a molecular
code, then the operator B(X) corresponding to it can be expressed by a
simple formula. (This is described in greater detail in Appendix 2.)
The boundary conditions, initial conditions, and laws of nature deter-
mine the density matrix p describing the physical state of affairs at time t . {
From p we can (at least in principle) calculate the probability, M(X), that
any particular molecule described by a code, X, can be found somewhere in
the system at time t :
x
M(X) = Trace(pB(X)) (10)

The probability function M(X) determines what molecular configurations
can be expected to have evolved in the system by the time t . If M(X) is1
fairly large, X can be expected to have evolved, but if it is very small the
evolution of X is unlikely.
According to our formal definition of information content, the informa-
tion content, L ( M ) , of the function M is equal to the length of the shortest
computer program that will calculate this function. This length is no greater
than the total length of the programming for all the various calculations we
have just described, and it should correspond to about three or four pages
of tightly packed programming instructions for a model based on an initial
primordial soup evolving in the presence of solar radiation by the laws
given in Figure 1. (We can envision this program as being written in a pro-
gramming language that provides for handling numbers with an arbitrary
number of significant digits. This language and the estimation of L ( M ) are
discussed in Appendix 2.)
We shall see in Section 5.3 that M(X) must be exceedingly small if
L(X), the information content of X, is very much greater than the informa-
tion content of M. The difference between the information in X and that
supplied by the system (represented by L(M)) must be made up by pure
chance, and therefore the probability of X goes down exponentially with
this difference. Thus the evolution of life forms in such a system is extremely
unlikely, for it can be argued that the length of the shortest program
needed to calculate the essential molecular structures of even a "primitive"
living organism should be a great deal longer than three or four pages. This
argument is discussed in detail in the next section.
Figure 2. The general form of a physical model.

We should note the ways in which "chance" enters into the function
M(X) for the probability of finding X in the system. There is an element of
randomness in the thermodynamic ensembles defining the initial condi-
tions and the boundary conditions. This is one source of randomness for
the "random" mutations of the theory of evolution, which are supposed to
be due to chance molecular collisions and bombardment by cosmic rays.
Also, randomness is built into the basic structure of quantum mechanics,
since the quantum mechanical states are essentially statistical in nature.
Thus the random mutations of the theory of evolution are automatically
built into our physical models. As we have pointed out before, natural
selection is also implicit in these models, for it is not an independent prin-
ciple, but must simply be a consequence of the underlying natural laws.
As a final point, we would like to mention another class of models that
exhibit the processes of mutation and natural selection involved in the
theory of evolution. These are the cellular automaton models pioneered by
12
John von Neumann. In these models an elemental finite-state automaton
is placed on each square of a large two-dimensional lattice. The state of the
system is given by specifying the states of all the automatons in the lattice.
The system changes in the following way with the passage of time: Let t be
a small, fixed time interval. At the end of each successive interval f, each
automaton changes to a new state in a way which depends only on the states
of the automatons occupying the immediately adjacent squares.
By properly adjusting the states of the automatons in a region of the lat-
tice, one can create an "organism" that functions through the interaction of
the automatons in its constituent cells. Von Neumann has shown that a self-
reproducing organism capable of exhibiting complex behavior (a universal
Turing machine) can be constructed in his cellular automaton model. His
idea was to prove that mechanical systems can exhibit the property of self-
reproduction characteristic of living organisms. This was intended as one
further step in the demonstration that life is a mechanical process governed
by mathematical laws.
It is therefore interesting to inquire whether self-reproducing "organ-
isms" of the kind von Neumann considered could evolve in his cellular
automaton system in a reasonable length of time. The system's law of trans-
formation with time should determine the natural selection of various forms
of organisms through the same sort of competitive processes envisioned in
ordinary evolutionary theory. The random mutations of evolutionary
theory could easily be introduced through a Markov process, which would
make random changes in the states of the elemental automatons.
One can imagine a scenario in which primitive self-reproducing organ-
The Great Complexity of Biological Form 111
isms could gradually arise through processes of self-organization in an ini-

tial "soup" consisting of automatons with randomly chosen states. These
organisms might then evolve sensory systems and arrangements for de-
fense and offense. Finally, intelligent beings might arise who would build
the equivalent of cities and empires, and create art, technology, and even
cybernetics.
But one could not expect such evolution to occur. Owing to the great
simplicity of von Neumann's system, the function corresponding to M(X)
for the system can be defined by a very brief list of equations. In contrast,
von Neumann's self-reproducing machines are very complicated, and he
required several hundred pages to describe them in his book. It therefore
seems reasonable to suppose that their information content greatly exceeds
that of M. As we shall see in Section 5.3, this rules out the evolution of such
configurations within the system.
5.2 The Great Complexity of Biological Form

In this section we shall discuss four aspects of the bodily structures of
living organisms: the molecular structures of cells, the genetic coding
(genotype), the visible bodily form (phenotype), and the behavioral
patterns of organisms. Under the last heading we shall be particularly
concerned with human behavior and its byproducts, including language,
literature, technology, and scientific theories.
First, let us consider the molecular structures of cells. One of the most
thoroughly studied organisms is the bacterium Escherichia coli, a uni-
cellular organism about 500 times smaller than the average cells of higher
13
plants and animals. This is one of the smallest and simplest of all living
organisms. Yet it is estimated that a single E. coli cell contains between
3,000 and 6,000 different types of molecules. Among these are some 2,000
to 3,000 different kinds of proteins, with an average molecular weight
of 40,000. James Watson, one of the foremost authorities on molecular
biology, admits that most of these large biological molecules do not obey
any simple structural rule:
Most of these macromolecules are not being actively studied, since their
overwhelming complexity has forced chemists to concentrate on relatively,
few of them. Thus we must immediately admit that the structure of a cell will,
never be understood in the same way as that of water or glucose molecules.
Not only will the exact structures of most macromolecules remain unsolved,
14
but their relative locations within cells can be only vaguely known.
Biochemists describe proteins as chains built up from twenty different

varieties of amino acid molecules. A typical protein molecule in an E. coli

• cell will contain some 300 of these amino acid subunits. Since each subunit
may be any one of twenty different amino acids, this means that the
300
number of possible protein molecules of typical size is about 2 0 . In gen-
eral, a binary number of log N digits can be used to identify a particular
2
member of an N-element list. It follows that no more than log N bits of 2
information are required to specify the member, given that the structure
of the list is known. (A bit is a binary digit of 0 or 1.) This comes to about
300 log 20= 1,297 bits for a typical protein molecule in an E. coli cell. Since
2
there are approximately 2,000 to 3,000 different proteins of this kind in a

cell, the total information content for cellular protein is bounded by an
upper limit of between 2,594,000 and 3,891,000 bits.
Each E. coli cell contains at least one chromosome, which consists of a
9
* circular molecule of DN A with a molecular weight of about 2.5 x 10 . This
D N A molecule is believed to contain coded instructions that determine
the structures of all the other molecules in the cell. D N A molecules can
be visualized as helical chains composed of successive pairs of D N A
bases. There are four different kinds of bases—adenine ( A ) , thymine ( T ) ,
guanine ( G ) , and cytosine (C)—and these are limited to forming four dif-
ferent kinds of base pairs: A - T , T - A , C-G, and G-C. Biochemists have
worked out a genetic code, whereby sequences of D N A base pairs repre-
sent sequences of amino acid molecules. In this code, each group of three
base pairs along the D N A chain specifies a particular amino acid in a cor-
responding protein molecule, or it codes for the termination of a protein
chain.
Each D N A base pair has a molecular weight of about 660, and so a group
of three pairs has a molecular weight of about 1,980. Since the molecular
9
weight of haploid E. coli D N A is about 2.5 x 10 , this means that the gene-
6
tic coding for this organism will consist of some 1.3 x 10 triplets of bases.
Since each triplet can discriminate between 21 alternatives (20 amino acid
6
types and a stop code), this gives us an upper bound of about 1.3 x 10
6
times log 21, or some 5.5 x 10 bits, for the genetic information content of
2
an E. coli cell. (In this section we carry out all calculations to several deci-
mal places, and then round off all numbers cited in the text to a few signifi-
cant figures.)
Yet the E. coli bacterium is a very simple organism. In the cells of higher
plants and animals, much larger amounts of D N A are found than in E. coli.
It is estimated that mammalian cells contain some 800 times as much D N A ,
9
yielding an upper limit of about 4.4 x 10 bits for the genetic information
15
content of these cells. Some idea of the size of these numbers can be ob-
tained by considering the number of pages of print required to write down
such large amounts of genetic coding in full. In a typical book there are
some 70 characters per line and 40 lines per page. This gives us about
4
1.7 x 10 bits per page if we use an alphabet of 64 characters. (Since
log 64=6, we can use 6 bits to encode each character.) At this rate it would
2
take about 330 pages to write down the coding for E. coli, and about
264,000 pages to write down the coding for a mammalian cell.
One of the dogmas of molecular biology (called the central dogma, in
fact) is that all the information needed to specify a cell is contained in the
cell's D N A coding, and that this coded information is not changed except
by random mutations. Of course, D N A cannot generate a functional cell
out of disorganized components by itself, and cells may therefore contain
structural information that is not coded in D N A . Also, the recent discovery
of reverse transcription shows that information can be transferred to the
16
D N A molecules from other molecules in the c e l l .
Nonetheless, the amount of D N A in cells should give us some idea of the
amount of information needed to adequately describe them. According to
the central dogma, the figures we have derived thus far should provide
upper limits for the essential information content of cells. We would now
like to estimate reasonable lower limits. To do this we shall need to con-
sider the variety and complexity of the structures cells are found to contain.
Since some genes in the chromosomes of higher animals apparently exist in
multiple copies, it would seem that the total genetic information content
for these organisms must be lower than the figures suggested by their total
D N A content. This conclusion is also indicated if, as some biologists have
argued, the chromosomes of higher organisms contain many stretches of
17
random, nonfunctional coding.
Yet based on the great complexity of structure visible under the micro-
scope in vertebrate cells, Watson estimates that such cells must be "at least
20 to 50 times more complex genetically than E. coli." By this he means
18
that such cells should contain coding for at least 20 to 50 times as many
kinds of protein molecules as E. coli. We would need between 6,600 and
16,500 pages to write down this amount of coding, based on our 330-page
8
estimate for E. coli. This corresponds to between 1.1 and 2.8 x 10 bits of
information, and between 84,000 and 210,000 proteins, with an average of
300 amino acid subunits apiece.
To determine a lower limit for the genetic information content of cells,
let us consider how information may be distributed among the protein
molecules of a typical mammalian cell. From the figures we have cited,
the lowest estimate for the number of distinct proteins in such a cell is
2,000 x 20=40,000. A protein of 300 amino acids can be specified by a string
of 1,297 bits, or binary 1's andO's . Since the size of proteins can vary, let us
assign a string of 10,000 bits for each protein. We shall label these strings
Y , . . . ,Y , where N=40,000, and the strings are listed in numerical
1 N
order. We can join Y , . . . ,Y together to form a large string, X, of

1 N
8
4 x 10 bits, and for each integer n = 1, . . . ,N we can form a string X of n
intermediate length by joining together Y , . . . , Y . 1 n
Figure 3. The structure of the bit string, X, representing the amino acid sequences
for the proteins of a typical mammalian cell.
The information content of X can be approximated by adding together

the amounts of information contributed by each of the blocks, Y „ , into
which X is divided. This idea can be formally expressed in the following
way. There is a special mathematical function, F, called the generating
function. This function can be used to calculate Y , . . . ,Y by means of 1 n
the formula,
Y n = F(w ,X )
n n-1 (11)
where n runs from 1 to N, and w , , . . . ,w are suitably chosen bit strings.
1 N
(Here we define X =1, so that the formula can be applied when n = l. The
0
mathematical terms and formulas we shall use are discussed in Appendix 1,

and we present only a brief introduction here.)
The idea of this formula is that Y, can be calculated from w . Then Y 1 2
can be calculated from w and Y =X . In general, each Y can be gener-

2 1 1 n
ated from w and the Y 's that have been already calculated. Each bit
n k
string w represents the information that has to be introduced to specify

n
Y , given that Y , . . . ,Y are known.

n 1 n-1
The w„'s are related to L(X), the total information content of X, by the
following inequality:
Here l(w ) represents the number of bits in the bit string w . This inequal-
n n
ity says, in essence, that the information content of X is at least as great as

the sum of the amounts of information introduced step-by-step as X is
assembled from Y . . . , Y .
1 N
5
Suppose for the sake of argument that L ( X ) 6 x l 0 bits. Inequality
(12) can be rephrased as
5 4
If we substitute 6 x 10 for L(X) and 4 x 10 for N, we find that the aver-
age value of l(w„) is no greater than 16 bits. This means that a typical
block Y„ can be calculated exactly from X using no more than 16 bits of
n-1
information.
Now, Y represents a protein molecule built from an average of 300
n
amino acids, and to directly spell out the amino acid sequence of such a pro-
tein requires some 1,297 bits. It is possible, of course, to specify some se-
quences of 300 amino acids with much less information. For example, if a
protein consists of a pattern of 10 amino acids repeated 30 times, then the
protein can be specified by about 43 bits of information. The amino acid se-
quences of actual proteins do not seem to follow any simple rule, however,
and it is therefore hard to believe that they could be consistently generated '
using such small amounts of information.
We can obtain an intuitive idea about the structural requirements of pro-
teins by considering the roles they play in the metabolism of cells. The
exact way these proteins function is far from being known at the present
lime. But it is known that they are able to behave like small computers.
19
Here is an example taken almost at random from Watson: In E. coli there
occurs a sequence of chemical reactions that convert the compound
threonine into isoleucine in five steps, each of which is catalyzed by a par-
ticular enzyme, or protein macromolecule. The first step in this sequence
can occur only with the aid of the enzyme threonine deaminase. When the
final product, isoleucine, has reached sufficiently high concentrations, it
interacts with the threonine deaminase molecules in such a way that they
no longer catalyze the first reaction of the series. This prevents the manu-
facture of more of the product than the cell needs. We may note that each
enzyme is so structured that it catalyzes only a few specific reactions. While
not affecting the rates of other chemical reactions at all, such biological en-
zymes are known for their ability to cause certain chemical reactions to
occur millions of times faster than they will occur under laboratory condi-
tions in which the enzyme is not present. We may also note that threonine
deaminase is inhibited specifically by isoleucine at the proper level of con-

centration, and not by any other chemical that would normally be present
in the cell (for otherwise this control system wouldn't work).
Molecular biologists hypothesize that many of the amino acids in an
enzyme are organized into "recognition sites" that respond to specific
molecules in the cell in the same way that a lock responds to a particular
key. The enzyme is so structured that its properties change in a systematic
way when it interacts with these molecules, and this enables it to act as a
logical unit in the overall functioning of the cell. According to this concep-
tion, an enzyme may be compared with a subroutine in a complex com-
puter program, and its constituent amino acids may be compared with
elementary programming operations that are combined to form a logical
network.
If we consider that 16 bits are not sufficient even to define an arbitrary
sequence of four amino acids, we can only conclude that most enzymes
cannot be independently specified by such small amounts of information.
Now, if one enzyme were very similar to another, it might be possible to de-
fine its structure by specifying small modifications in the structure of the
other enzyme. This would make it possible to calculate the Y for this en- n
zyme using the Y - for the other enzyme plus a small amount of additional
n
information. Yet if we examine the structure and function of cells, we find

that they contain many diverse and complex enzymes. There is no reason
to suppose that these enzymes can be systematically transformed from
one into another using on the average only 16 bits of new information per
transformation.
Also, there is no reason to suppose that the generating function, F, has
unique properties that enable it to produce the amino acid sequences of
proteins. After all, F is determined by a simple mathematical definition
that has nothing to do with living organisms. (F is defined in Appendix 1.)
We conclude that we must be incorrect in assuming that the informa-
5
tion content, L(X), of mammalian proteins is less than 6 x 10 bits. We
therefore propose that this figure can be taken as a reasonable lower bound
forL(X).
5
Lower bound of L ( X ) > 6 x l 0 bits (14)
This conclusion is given further support if we perform the calculation in for-
mula (13) using a value of N=210,000 corresponding to our upper estimate
for the number of distinct proteins in mammalian cells. Using this N, we
5
find that l(w ) 4 bits on the average if we assume that L(X) 6x 10 bits.
n
1 9
We can even go further and let A designate a string of 4.4 x 10 bits rep-
resenting the complete genetic coding of a higher plant or animal. As we
have already mentioned, this X corresponds to 264,000 pages of coding at

4
1.7 x 10 bits per page. I f w e l e t N=264,000 and let Y . . . , Y represent
1 N
20
these pages, then we can again perform our calculations. We find that if
5
L ( Y ) 6 x 1 0 , then the average l(w ) must be less than 4 bits. This implies
n
that, on the average, each successive page of the mammalian genetic cod-
ing introduces no more than 4 bits of new information. In other words, the
typical page Y is given by F(w,X ), where w is an integer between 1 and
n n-1
15, and X n-1 represents the preceding pages. This strongly suggests that
5
L ( X ) > 6 x l 0 f o r this X.
At this point the objection might be raised that while A"must indeed have
a high information content, most of this information may well consist of
insignificant nonsense. One reason for supposing this is that molecular
biologists have discovered in the cells of higher organisms sections of D N A
coding that serve no obvious purpose. These sections of coding can pre-
sumably change freely under the influence of mutations, and thus they may
be expected to consist largely of random noise.
This objection can be answered by pointing out that while such random
noise may be present, the structural and functional requirements of living
organisms clearly call for large amounts of meaningful information. This
was the basis for our argument that the information content of mammalian
5
proteins must be at least 6 x 10 bits. In addition, biologists generally find
that structures in living organisms have some important function, even
though it may not be immediately obvious what this function is. This
suggests that most of the D N A coding in the cells of higher organisms
should contain some significant information.
We can conclude that our estimates of minimal information content re-
flect the presence of significant information, but that the strings, X, to
which these estimates refer may also contain some irrelevant noise. We
shall therefore introduce the idea of a symbolic description that represents
essential biological information, while omitting irrelevant details. Our
string of protein sequences is an example of such a description, although it
may have the shortcoming of including some insignificant features of pro-
teins. In general, a symbolic description is a bit string that encodes biologi-
cal information, but which does not necessarily correspond directly to the
genetic coding of an organism. Such a description can omit random noise
that may be present in the genome, while at the same time including an
account of important features of the organism.
At the present stage of biological knowledge, it is not possible to define
in detail an adequate symbolic description of a cell. Yet if a symbolic
description could be formulated that extracted only 24 bits of significant
information from each (17,000 bit) page of the mammalian genome, then
6
the information content of this description would be at least 6 x 10 bits.
To give a further indication of the complexity of living organisms, let us
now describe a number of different categories of organic structure. The
structures in each of these categories could, in principle, be formally repre-
sented by symbolic descriptions that capture their essential features. Since
these structures tend to be highly complex and variegated, we suggest that
their corresponding formal descriptions must have a large information
content.
We shall consider the following hierarchy of structure and function in
living organisms.
(a) The chemical reactions involved in cellular metabolism. These in-
volve respiration, the synthesis of various chemicals needed in the cell from
food molecules, photosynthesis in plants, and the processes involved in the
orderly breakdown of different molecules. It would appear that most of the
genetic coding of E. coli must be devoted to metabolism, since these bacte-
rial cells do very little but grow and divide in half. Even though E. coli is
one of the simplest of organisms, its metabolic interactions are so intricate
that "the exact way in which all these transformations . . . occur is enor-
mously complex, and most biochemists concern themselves with studying
21
(or even knowing about!) only a small fraction of the total interactions."
That these interactions must be governed by a complex system of logic
rivaling the most sophisticated programs of modern electronic computers
is certainly indicated by the descriptions in Watson's book.
(b) The morphology of cells. The E. coli cell appears to possess a rela-
tively simple gross structure, but many cells, even among the algae and
protozoa, have a very complex morphology. For example, the cilia of
protozoa such as the paramecium have been shown to possess an intricate
structure. The cilia function cooperatively under the direction of versatile
22
cellular control mechanisms to produce a synchronized rowing machine.
They are assembled within the cell according to a regulated program, and the
process of construction must also require complex molecular machinery.
In the cells of higher organisms, we find many examples of complex
morphology. For example, biologists have discovered several kinds of
23
structures that join cells together and enable them to communicate.
These junctions display an elaborate architecture of sheets, tubes, and fila-
ments, and some of them are designed to open and close systematically in
response to cellular conditions. It seems very doubtful that precise plans
for this cellular machinery could be specified by a small amount of informa-
tion. It also seems doubtful that these plans could be generated by adding a
lew bits of information to, say, the plans for the enzymes regulating the
Krebs cycle of cellular respiration.
(c) The varieties of cells making up the tissues of higher organisms. We
can easily write down a long list of types of cells appearing in diverse bodily
organs. These include muscle cells, nerve cells, bone cells, different kinds
of blood cells, glandular cells, liver cells, epithelial cells, and so forth. The
study of a particular kind of cell can be a whole academic subject in itself,
and doctoral dissertations are frequently devoted to the study of a detail of
a detail of the structure and function of such cells. The complete instruc-
tions for constructing all these cells must be contained in the genetic coding
of any higher organism, at least according to the understanding of modern
biology. There must also be instructions controlling the development of
these cells during the growth of the embryo.
(d) The structure and function of the organs in higher plants and animals.
The various organs of the body perform a vast array of complicated func-
tions, most of which are poorly understood. Examples include the disease-
lighting system of the blood, the image-producing eye and its retina, the
brain, the endocrine gland system, and the heart and circulatory system.
The functions of a typical organ are carried out through the cooperative
interaction of many complex subsystems. For example, the subsystems of
(he eye include the lens, the muscles supporting it, the iris, the cornea, the
retina, the nerve connections, and the muscle system that moves the
eyeball. Many of these subsystems are very intricate. For example, the iris
contains a muscle system for opening and closing the pupil; the lens must be
transparent and shaped so as to focus a sharp image on the retina; the retina
contains systems of cells and nerves designed to detect elementary visual
patterns such as lines and edges; the light-sensitive cells contain complex
chemical systems designed to respond to different colors of light; and so
forth. It would not be surprising if significant amounts of information were
required to represent each of these subsystems as combinations of their
underlying cellular components. In general, we would expect the basic
plans of organs to have a high information content.
(c) The behavior of animals other than man. Many patterns of complex
behavior are exhibited by lower animals. We may note, for example, the
social systems of bees and ants, the spinning of spider webs, and the trans-
continental migrations of birds. Biologists generally think that these instinc-
tive behavioral patterns are built into the genetic material of the organisms. -
This suggests that they must be represented biochemically as a series of
logical " i f - t h e n " instructions similar to the program of a computer.
It might be an interesting challenge for a student of animal behavior to

- try to write computer programs that would duplicate the behavior of a
particular animal. As an indication of the difficulty this would involve,
consider the problem of pattern recognition. Many birds are highly
, discriminating in their response to the coloring and physical shape of other
birds. This means that they are able to make fine distinctions between com-
plex patterns of color and form. Yet computer programs that can discrimi-
nate between simple geometric shapes have proven very difficult to write,
and have involved many elaborate programming procedures. (For exam-
6
ple, one such program, called the " M I T robot," requires some 3.6 x 10
24
bits of programming instructions. ) We can conclude that the coded in-
structions specifying animal behavior may be expected to entail a great deal
of information.
(f) Human personality. This is perhaps the most complex topic of all,
and it is presently beyond the reach of the reductionistic methods of mod-
ern science. There are substantial reasons for believing that the conscious
self cannot be understood in physical terms, and this suggests that the
behavior of conscious beings also cannot be fully explained in terms of
physical processes. Nonetheless, scientists who accept the mechanistic
world view are committed as a matter of consistency to suppose that all
. aspects of personality can be described in terms of interacting molecules.
For the sake of argument, let us consider the hypothesis that personality
can be adequately defined in this way. Our question is: How much informa-
tion would be required to write down a symbolic description capturing the
essential features of human personality?
Such a symbolic description would have to define human intelligence,
and it would have to explain the faculty of speech and the abilities involved
in artistic and musical creativity. In addition, this description would have to
represent the basic qualities that characterize interpersonal behavior. The
following list of personal attributes may give some idea of what this entails:
Anger, anxiety, apprehension, arrogance, bashfulness, conceit, compassion,

courage, determination, doubt, dread, envy, faithfulness, forbearance,
forgiveness, gratitude, gravity, greed, guilt, haughtiness, humility, impu-
dence, lamentation, mercy, patience, peacefulness, perseverance, pride,
renunciation, respectfulness, tactfulness.
We may obtain some idea of the amount of information needed to define

these attributes by considering the great diversity and complexity of human
artistic and literary productions. Of course, it is not possible at the present
time to distill a precise characterization of personality from this source of
raw data. Yet we wonder whether the information content of a full sym-
Limitations on the Evolution of Complex Form 121
bolic description of human personality could be as low as our minimum

estimate of the information content of a higher cell. We recall that this was
5
6 x 10 bits, or about 36 pages.
5 . 3 I n f o r m a t i o n - t h e o r e t i c Limitations
on the Evolution of Complex Form
In Section 5 . 1 , we discussed the mathematical models of natural phe-
nomena used in modern physics, and pointed out that the fundamental
program of the physicists is to make these models as simple and general as
possible. The basic format for a physical model is as follows: We have a sys-
tem, 5, in which events take place in accordance with a set of natural laws,
F. Generally, F will be defined by a set of differential equations. The sys-
tem S will have initial conditions, S , and boundary conditions, B. These
0
will be described by probability distributions which may consist of combi-

nations of thermodynamic ensembles such as the Gibbsian canonical en-
semble. The probability distribution, P , for the state of the system at time t
1
can be calculated, at least in principle, from the equations for the initial
conditions, boundary conditions, and natural laws.
We have argued that simplicity in a mathematical model can be mea-
sured by the amount of coding needed to express the model in a fixed pro-
gramming language. The mathematical models of physics tend to be simple
in the sense that they can be specified by relatively small amounts of cod-
ing. For example, the Schr dinger equation of Figure 1 can be solved nu-
merically by a program occupying less than one solid page of programming
instructions. (This is discussed in greater detail in Appendix 2.) If we allow
four additional pages for initial and boundary conditions, we find that we
can express in less than five pages a model of a primordial planet evolving
in accordance with the laws of chemistry and electromagnetism. The infor-
mation content of P, for this model is no more than 14,000 characters (at
2,800 characters per page), or
L(P )
t 6x 14,000 = 84,000 bits (15)
(Recall that each character in our 64-character alphabet requires six bits.)
From the probability distribution P we can calculate the probability,
t
M(X), that a particular molecular configuration described by a code, X,

will be found somewhere within the system at time t. M(X) can be inter-
preted as the probability that the configuration X will have "evolved"
within the system by the time t. To calculate M(X) given A"and P we need t
a function that can analyze the state of the system at time t and report
whether or not the configuration represented by X is present. One can
define an adequate function of this kind in less than one page of condensed
programming instructions. (Again, see Appendix 2.) This gives us an upper
bound for L(M) of less than six pages, or
Upper bound of L(M) 100,800 bits (16)

In contrast, in Section 5.2 we estimated a lower bound for the information
content of a configuration X representing the genetic coding for a cell of a
higher organism. This estimate was:
Lower bound of L(X) > 600,000 bits (17)
The following inequality reveals the implications of these bounds for the
theory of evolution:
C + l o g 2 T + L ( M ) - L ( X )
M(X) 2 (18)
where c is a constant of 326 bits, and

T = The maximum number of distinct configurations, X, that (19)
can coexist within any one state of the physical system.
The molecular configurations should be defined in such a way that no two

distinct configurations share atoms in common. (For example, a part of a
configuration should not count as a distinct configuration.) Then Tis no
greater than the total number of atoms in the system. For a system the size
of the earth, this should be no more than T = 1 0 . Some physicists have 5 1
estimated that the total number of subatomic particles in the universe is

25
less than 1 0 . Using this speculative figure, we can estimate log T 266
80
2
in a universal model.
Putting these values together, we find that
3 2 6 + 2 6 6 + 1 0 0 , 8 0 0 - 600,000 -150,000
M(X) 2 10 (20)
According to inequality (20), the probability that the configuration X will

be found anywhere within the system at time t can be no greater than one
out of 10 . Suppose that t is 4.5 billion years—the estimated age of
150,000
the earth according to present geological theories. Multiplying M(X) in in-

equality (20) by 365 x 4.5 x 10 , we see that the probability of finding X
9
in the system at the end of any day in a 4.5-billion-year period is also much
-150,000
less than 10 . We can conclude that the entire course of events in S
over a 4.5-billion-year period would have to be repeated over and over
again at least 1 0 times for there to be a reasonable expectation that
1 5 0 , 0 0 0
the configuration X would be seen in 5 even once. (This is in accordance

with what is known as the frequency interpretation of probabilities. See
Chapter 6 for a discussion of probability and its interpretation.) Since
150
1 0 ' is an enormous number, to say the least, it is reasonable to suppose
that X will not evolve within the system in 4.5 billion years.
Now, X could represent the genetic coding for a higher animal, and 5
could be a mathematical model of an earth-sized physical system in which
the evolution of life might be expected to occur. Our conclusion is that
higher life as we know it will not evolve in such a system in any realistic span
of time. For the time t we could have chosen any time period for which L(t)
is small, for the only property of t that affects our considerations is the
number of characters needed to express t in the program for calculating P,.
This means that our t could be anything from one year to billions of billions
of years, and the conclusion expressed in inequality (20) would still hold.
Figure 4. A pictorial representation of inequality (18). The horizontal triangular

area represents the set of all possible biological forms, arranged so that their infor-
mation content increases from left to right. ( A s we go from left to right these forms
become more densely packed since the number of possible forms of information
content L(X) increases exponentially with L(X).)
The vertical axis represents probability. Peaks ( A ) and (B) represent proba-
bility distributions for the forms generated by a physical process with information
content L(M). The curved surface represents the restriction on probabilities im-
posed by inequality (18). The distribution (B) of complex forms violates this restric-
tion and is therefore ruled out. In contrast, inequality (18) does not rule out the
distribution ( A ) of simple forms.
It is often said that while it is highly unlikely that any given complicated
structural part of a living organism will arise by chance, still, given the
immense span of geological time, such organisms are bound to evolve
eventually. Self-reproducing systems of molecules will automatically arise
through the effects of random combination and processes of chemical self-
organization. These systems will be modified by random mutations, and
they will gradually evolve through the culling action of natural selection.
Yet, although this evolutionary picture may seem superficially plausible,
the analysis presented here rules it out in a system characterized by simple
natural laws and simple initial and boundary conditions. The reason for this
is easy to understand. Natural selection is supposed to direct the process of
evolution, but its only source of guidance is the information built into the
natural laws and the initial and boundary conditions of the system. This
means that natural selection can constrain the processes of random combi-
nation and mutation only in a simple way. Natural selection may be able to
impose simple patterns on the welter of randomly distributed molecules,
but in a simple model, complex order can only arise by chance.
For this reason, the theory of evolution can be seen to fail at the very
weak point that has been criticized by so many students of the theory.
Many observers have noted that natural selection has never been ade-
quately defined in evolutionary theory, either in Darwin's original version
26
or in the more recent "synthetic t h e o r y . " According to Darwin, natu-
ral selection means "survival of the fittest," but no one, unfortunately,
has been able to define which creatures are fittest, except by saying that
they are the ones that survive. A similar problem plagues the more recent
definition of natural selection as "differential reproduction."
Evolutionists have always had an intuitive feeling that the ordinary inter-
actions between living organisms would generate forces of selection which,
over long periods of time, would transform simple molecular combinations
into higher life forms. This intuition is based on commonplace examples.
For example, a mutation producing longer legs in a deer might well be
selected because it would enable the deer to escape from predators more
easily. Yet, this is no reason to believe that natural selection would possess
the discriminating power needed to guide the development of a world of
plants and animals from an inanimate primeval slime. What we have shown
here is that in a system governed by simple natural laws, no process is suffi-
cient to do this, whether it be natural selection or any other imagined
process of evolutionary development.
At this point, let us consider one possible objection that could be raised
to our interpretation of inequality (18). We have discussed the question of
whether or not a configuration represented by X could evolve within the
system S. But do the X's we have considered represent objects of real inter-
est? Suppose, for example, that X represents the exact D N A coding of a
particular cell. Many mutations are believed to have a neutral effect on
cellular development, and many others might affect the cell in a relatively ,
insignificant way. By combining, say, 10,000 mutations of this type in vari-
1 0 0 0 0
ous ways, one could conceivably produce 2 different cells, all nearly
identical in form and function. One might argue that while the evolution
of any one of these cells is highly improbable, it might nonetheless be prob-
able for at least one out of the total collection to evolve. If p is the proba-
bility for the evolution of each individual genotype, then the probability
for the evolution of at least one of the genotypes might be in the order of
2 w s m
p . This probability could be fairly large, even if p were very small.
Perhaps the most comprehensive way to answer this objection is to make
use of the concept of symbolic descriptions that we introduced in Section
27
5 . 2 . Recall that a symbolic description of an organism is a sequence of
coding that specifies significant features of the organism, while omitting
unimportant details. We argued in Section 5.2 that an adequate symbolic
description of a higher organism should possess a high information content,
and we argued that the same could be said of the symbolic descriptions
defining many important general features of living organisms.
The following diagram sums up the relationship between a symbolic
description and the many individual instances to which it refers.
Here X , . . . ,X represent distinct configurations of matter, and Y repre-

1 n
sents a symbolic description of these configurations. G designates an

"observation function," or a process of analysis that can be applied to any
of the X 's to yield Y.
j
Computer programs for pattern analysis provide many examples of ob-

servation functions. For example, consider a program for the analysis of
handwriting. If X , . . . ,X were different handwritten copies of a par-
1 n
ticular text, this G could produce a printed version of the text by analyzing
any one of the copies. The printed text Y would contain the essential infor-
mation in X , . . . ,X , while omitting all reference to the individual styles
1 n
of handwriting they might display. We can think of Y as a symbolic descrip-

tion that applies equally well to each of the handwritten copies, even
though these copies may differ greatly from one another in detail.
In the field of biology there are many applications for the concepts of
symbolic descriptions and observation functions. For example, consider a
particular species of animal, such as the horse. Horses vary greatly in indi-
vidual detail. But if there is any meaning to the term "horse," there must be
certain specific information that is common to all normal horses and that
fully defines this category of organic form. It should be possible to ascertain
this information by observing any horse that is not defective in some way,
for otherwise it could not be said that the information truly characterizes
horses. This suggests that there should exist a symbolic description charac-
terizing horses, and an observation function G that can analyze any given
horse and generate this symbolic description.
In general, an observation function can be constructed from any sys-
tematic process of observation that abstracts certain features from ob-
served objects and ignores others. One of the principles of reductionistic
science is that all such processes can be defined mathematically as al-
gorithms. Of course, no one has yet devised algorithms for the kind of
sophisticated observational processes that we are considering here. In this
chapter, however, we will assume for the sake of argument that this is
possible, and we will treat observation functions as computer programs.
(A technical discussion of observation functions is given in Thompson,
(1980).)
If Y is a symbolic description produced by an observation function G,
we can measure the information content of Y by L(Y | G ) . The symbol
L(Y |G) designates the amount of information required to specify Y, given
that G is known. The idea is that we only want to measure information
in Y that is obtained from X , . . . ,X . We do not want to consider any
1 n
information in Y that is contributed by the observation function itself.

We can clarify this idea by referring once again to the example of hand-
writing analysis. In this example, Y might consist of the original text
printed in a particular typeface. In that case the information defining this
typeface could be ascertained by inspecting Y. Yet this information has
nothing to do with the original handwritten texts, X , . . . ,X , and is sim-
1 n
ply a byproduct of the observation function, G. The quantity L( Y \G) rep-

resents the amount of information in Ythat is not derived from G, and thus
L ( Y |G) represents information that is coming solely from X , . . . ,X . 1 n
Using the concept of observation functions, we can generalize inequality

(18) in the following way:
C + l o g 2 T + L ( M ) - L ( Y |G)
M(X )
1 + M(X ) + . . . + M(X )
2 n 2' (22)
Here c'=446 bits, and T is the same as before. Y represents a symbolic

description of organic form, and G represents an observation function.

X , . . . ,X represent all the individual configurations of matter to which
1 n
G assigns the description Y. For example, Y might be a symbolic descrip-

tion of ahorse, and X , . . . ,X might represent all the possible individual
1 n
horses satisfying this description.

We have argued in Section 5.2 that an adequate symbolic description of,
5
say, a horse must have an information content of 6 x 10 bits at the very
least. (Recall that such a description must define from scratch all signifi-
cant structural features of the horse—from cellular organelles to gross
anatomy.) Since this information should be independent of the function G
used to observe it, we would expect to find L(Y |G) 6xl0 bits. We can 5
thus recast inequality (20) in the form,

-150,000
M(X )
1 + M(X ) + . . . + M(X )
2 n << 10 (23)
Here the summation, M(X ) + . . . + M(X ), is an upper bound on the

1 n
probability that some one of the forms X , . . . ,X will evolve within the
1 n
-150 0 0 0
system S. Our conclusion is that since 10 is an exceedingly small
number, it is nearly certain that none of these forms will ever evolve in S. In
the particular case we are considering, this means that no instance of a
horse will ever be found in 5. In general, it means that no form with a sym-
bolic description of high information content can be expected to evolve in a
system characterized by natural laws, initial conditions, and boundary
conditions of low information content.
Let us examine inequality (22) more closely to understand the basic prin-
ciple behind it. First of all, note that the upper bound this inequality places
on M(X ) + . . . + M(X ) will become reasonably large only if we increase
1 n
the complexity of our model to the point where L(M) is nearly equal to
600,000 - 266 - 446 bits, or nearly 36 solid pages of coding. There are
three ways of increasing L(M). One way is to increase the complexity of the
laws of the system. We note, however, that scientists have not yet reached
the point of even considering natural laws requiring 36 pages of coding, and
it is unlikely that such considerations could ever be practical, given the
limitations of the human mind. At any rate, the laws of present-day phys-
ics are based on the Schr dinger equation and a few potentials, and the
fundamental program of physics has been to simplify these as much as
possible.
L(M) might also be increased by increasing the complexity of the initial
conditions or the boundary conditions of the system. As we shall see, this
would amount to building data for the symbolic description, Y, into the
initial or boundary conditions. If specifications for Y were built into the
boundary conditions, then as time passed we would find information for Y
coming into the system from across the boundary. If such specifications
were built into the initial conditions, it would appear that information for Y
had existed in the system from its very beginning. In either case, the mathe-
matical model provides us with no satisfactory explanation of the original
source of the information for Y. Certainly neither alternative is acceptable
from the point of view of current evolutionary theory.
We can clarify these points by considering the following stronger version
of inequality (22):
C + log2T - L(Y |G,M)
M(X )
1 + . . . + M(X ) n 2' (24)
Here L(Y \G,M) designates the amount of information needed to specify

y, given that both G and M are known. This is formally defined as the
length of the shortest program that can calculate Y, using the functions G
and M as built-in subroutines.
If L(Y |G,M) is small, most of the information for Y can be obtained
from G and M. We have already postulated that L(Y |G) 6xl0 bits. In 5
order for L(Y |G,M) to be much smaller than this, it is necessary to adjust
M very carefully and, in effect, build specifications for Yinto M. Inequality
(24) shows that there can be a reasonable probability of finding an X in the j
system at time t only if L(Y |G,M) is not very much greaterthanc' + log T. 2
This constant is no more than 712 bits, or about 119 characters. This means
that nearly all the information needed to specify Ymust be built into M in
the form of natural laws, boundary conditions, or initial conditions.
This is an appropriate time to consider one objection that might be made
to our arguments. One might point out that the information content of or-
dinary matter is certainly very high. For example, the information needed
to specify the positions of the billions of molecules in a gas must be enor-
mous. How, then, are we justified in saying that L(M) for a physical model
should be small? The answer to this question is that we have described the
initial state of our physical system by means of thermodynamic ensembles
that specify only broad statistical properties of matter. However, we could
also consider a model in which the initial conditions stipulated a particular
arrangement of subatomic particles in a primordial gas.
In such a model, L(M) would be extremely large, and inequality (22)
would not tell us very much. But what should we expect of L(Y |G,M) in
this situation? There are essentially two alternatives. If the arrangement of
our primordial gas were chosen at random, then we would expect the infor-
mation specifying this arrangement to be irrelevant to the construction of
Y, and thus we would expect L(Y |G,M) to be large. According to (24), this
would mean that M{X ) + . . . + M(X ) would be very small. (This line
1 n
of reasoning is automatically incorporated in the derivation of (22) and

(24) if we use statistical ensembles in the definition of M.) Conversely, if

L(Y |G,M) were small, specific information for Y would have to be built
into the initial distribution of particles in the primordial gas. One would
then be led to ask how this information came to be encoded in such a
strange form in the beginning of the history of the physical system.
The same considerations apply to boundary conditions. For example,
if cosmic rays entered the system from outside, and their energies and
directions of motion were given by a simple statistical distribution, then the
system could not be expected to extract a large amount of specific infor-
mation from them in the form of an organism with a symbolic description,
Y. In order for this information to be obtainable from the cosmic rays,
their statistical distribution function, D, would have to be highly complex.
Furthermore, it could not be just any complex function, but would have to
be chosen so that L(Y | G , D ) was small. In other words, a large amount of
specific information for Y would have to be coded into the cosmic ray,
stream.
This brings to mind the panspermia hypothesis of Svante Arrhenius. Ac-
cording to this theory, life never had an origin, and it has been dissemi-
nated throughout the universe in the form of spores that travel in outer
space from one favorable planet to another. Arrhenius thought these
spores would be primitive and would independently evolve into higher
forms on various planets. We can see, however, that "primitive" spores
would not do—the spores would have to contain instructions for all of the *
organisms that later evolved.
The theory of Arrhenius holds that life is eternally existent as a material
phenomenon. This theory suffers from the drawback that natural processes
tend to obliterate information, thereby making it very unlikely that a large
store of information could be maintained eternally in a material system.
Also, the theory is no more satisfying as an explanation of origins than a
modified "big bang" theory, with instructions for life forms encoded in the
initial state of the universe.
Yet in a system having a low value of K=L(M) + l o g T + c ' , the proba-
2
bility is practically zero that life as we know it will arise. This K can be
regarded as a universal constant of the physical system that constrains its
evolutionary potential. If any category of form can be characterized by a
simple description having information content less than K, then it is possi-
ble for representatives of that category to arise in the system. But if the
members of the category satisfy a description of higher information content
than K, then they are highly unlikely to evolve.
This applies to categories such as species, classes, and phyla of plants and
animals. It also applies to more abstract categories such as intelligence or
personality. We can conclude, for example, that personality will not evolve
in a system of low l v a l u e unless we define personality so broadly that it be-
comes practically devoid of meaning. This can be appreciated if we con-
template the list of features of personality given in Section 5.2. Unless
these features can be adequately defined by a description of low informa-
tion content, we cannot expect to see entities that display them arising in
a system of low information content. (Here, of course, we are speaking
of personality as a purely physical phenomenon, in accordance with the
reductionistic philosophy of modern science.)
At this point we would like to briefly discuss some recent theories of the
self-organization of matter. These theories are based on nonequilibrium
thermodynamics and the theory of chemical kinetics, and they have been
28
developed by researchers such as Ilya Prigogine and Manfred Eigen. In
these theories, systems of chemicals are modeled with differential equa-
tions of a kind known as reaction-diffusion equations. By studying such
equations, the theoreticians hope to show how disorganized molecules in a
primordial soup could combine together chemically to produce the first
self-reproducing cell.
Perhaps the most well-known example of this work is Eigen's theory of
29
self-reproducing hypercycles. In this theory, Eigen proposes that systems
of polypeptides and self-replicating R N A molecules will arise sponta-
neously in a suitable chemical soup. A collection of R N A molecules,
R1,...R , will be able to perpetuate itself in the soup if each ft generates
n y
polypeptides that catalyze the reproduction of R , and R similarly cata-

j+1 n
lyzes the production of R . Eigen calls such systems of molecules "hyper-

t
cycles," and he proposes that they might have filled the evolutionary gap
between disorganized chemical compounds and living cells.
Here we will not attempt to discuss Eigen's theory in detail. We will sim-
ply point out that Eigen has not even begun to explain the origin of the self-
reproducing machinery of living cells. In existing cells, the D N A molecules
carrying genetic information are reproduced with the aid of highly complex
and specific enzymes, such as " D N A polymerase" and " D N A gyrase." These
enzymes all perform functions which are apparently necessary for cellular
reproduction, and which frequently involve remarkable feats of molecular
manipulation. For example, recent work suggests that D N A gyrase is able
systematically to break a D N A strand, pass a nearby strand of D N A
30
through the break, and then rejoin the broken strand. Since long D N A
chains automatically become tangled during replication, this capacity to tie
and untie knots in D N A appears essential for cellular reproduction.
A l l these enzymes are manufactured within the cell by an elaborate pro-
cess requiring a large number of proteins and other complex molecules.
Complex Form and the Frustration of Empiricism 131
For example, one step in this process is carried out by structures called
ribosomes that in Escherichia coli have a molecular weight of about 3 mil-
31
lion, and can be broken down into 55 distinct protein subunits. A l l these
proteins are themselves coded in D N A , and they are manufactured by the
same process in which they play an essential role.
Eigen's problem is to show how a stable self-reproducing system could
begin to operate before the development of this elaborate machinery, and
then show how the machinery could gradually evolve by mutation and
32
natural selection. Dixon and W e b b summarize some of the many dif-
33
ficulties that must be overcome in such an attempt, and S m i t h gives
a critique of Eigen's theory. Here we will simply observe that Eigen's
theoretical framework is not adequate for his task. In a model based on
reaction-diffusion equations, each chemical compound is represented by a
continuous function describing its distribution in space and time. In such
models there is no direct way to describe three-dimensional geometrical
structures composed of molecules, even though such structures are very
important in living cells.
For this reason, evolutionary models based solely on reaction-diffusion
equations are necessarily incomplete. Such models cannot readily simulate
processes involving such three-dimensional structures as ribosomes or cell
membranes, and they are even less suited to showing how such structures
may have evolved. At present, there seem to be no mathematical models of
biochemical processes that are both reasonably complete and easy to han-
dle. Hence we have based the analysis in this chapter on the laws of quan-
tum mechanics, even though these laws are difficult to apply in practical
calculations, and are also subject to a number of theoretical drawbacks.
As a final point we note that an analysis of evolution on the basis of in-
formation theory has been carried out independently by Hubert Yockey.
Yockey considered the problem of chemical evolution in a primordial
soup, and he has given estimates for the information content of proteins
34
such as cytochrome-c.
5.4 Complex Form and the Frustration of Empiricism

In this chapter we have used information theory to study the hypothesis
that life has originated and evolved through the action of natural processes
governed by the known laws of chemistry and physics. There are a number
of ways of interpreting this analysis. Some of the more important of these
are the following:
(1) The universe is characterized by simple natural laws and simple ini-
tial conditions. Adequate descriptions of living organisms do require
large amounts of information, and the evolution of such organisms is in-

deed exceedingly improbable. Yet this is what happened—simply by
chance.
(2) Adequate descriptions of organisms have a high information
content, and this information is built into the boundary conditions or
initial conditions of the universe. Evolution is thus a process whereby
pre-existing information is transformed from one form to another.
(3) Considerable amounts of information are required to describe

living organisms, and this information is built into the fundamental laws
of the universe. These laws correspond to the absolute reality underlying
the flux of transient phenomena, and thus life, or at least the information
for life, can be regarded as absolute.
(4) Detailed descriptions of living organisms may seem highly com-

plex, but actually they have a low information content. We cannot see
how to define such descriptions with simple programs, but this is nonethe-
less possible, and thus evolution may take place with high probabilities in
simple systems.
(5) It is hopeless to speculate on these topics, for we cannot arrive at

any satisfactory or reliable conclusions.
Let us begin by discussing interpretation (1). Evolution is often pre-
sented in popular accounts as an inexorable, progressive development
that advances inevitably from primitive to advanced forms. Evolutionists
frequently emphasize that chance enters the theory of evolution only as
a source of necessary variations, and that the deterministic mechanisms
of natural selection provide systematic guidance for the development of
35
species. Intelligence in particular is often seen as an inevitable product of
evolution, and many scientists, such as the astronomer Carl Sagan, specu-
late on the possible development of human or superhuman intelligence on
36
other planets.
Yet, a number of evolutionists have expressed a contrary view. The evo-
lutionary theorist Stephen J. Gould has recently popularized the idea
that sheer randomness plays an important part in the origin and develop-
37
ment of l i f e . This point had been elaborated previously by Theodosius
38
Dobzhansky and George G. Simpson, who both maintained that the
evolution of man or manlike beings has a probability of zero. Simpson's
arguments are as follows:
The factors that have determined the appearance of man have been so ex-
tremely special, so very long continued, so incredibly intricate . . . and every-
thing we learn seems to make them even more appallingly unique. If human
origins were indeed inevitable under the precise conditions of our actual his-
tory, that makes the more nearly impossible such an occurrence anywhere
else. I therefore think it extremely unlikely that anything enough like us for
39
real communication of thought exists anywhere in our accessible u n i v e r s e .
Simpson's reasoning is based on his intuition that natural selection is not an

independent guiding force, but merely a collocation of innumerable special
factors that just happen, in this case, to produce human beings when added
together. Since he can see no reason for the pre-existence of these par-
ticular factors, Simpson concludes that our existence is an irreproducible
consequence of sheer chance.
Simpson's ideas about the improbability of human evolution seem com-
pletely incompatible with the general understanding of evolution as a pro-
gressive process. Of course, one might try to reduce this incompatibility by
limiting Simpson's observations to man, while maintaining that higher life
in general is still likely to evolve. Yet the analysis given here implies that
Simpson's conclusions must apply to all features of life characterized by
descriptions of high information content. This includes everything from the
complex structures of cells to highly abstract aspects of human personality.
In a system operating under simple physical laws and initial conditions,
such patterns of organization can arise only with a probability of nearly
zero. (See Section 5.3.)
Now, one might assert that this is exactly what has happened. This is in-
deed the message conveyed by Jacques Monod in his book Chance and
Necessity:
The thesis I shall present . . . is that the biosphere does not contain a predict-
able class of objects or of events but constitutes a particular occurrence, com-
patible indeed with first principles, but not deducible from these principles
40
and therefore essentially unpredictable.
Monod is proposing that life in all its diverse manifestations constitutes a

unique phenomenon that has arisen simply by chance. Monod believes, of
course, that life follows physical laws, and that organisms must be con-
structed in accordance with the dictates of these laws. (Thus land animals
cannot exceed a certain size, for example.) But apart from this he regards
life as completely unpredictable.
Monod's view is consistent with the analysis presented here, but this
view is the antithesis not only of the theory of evolution, but of any effort
to scientifically explain the origins of life. In effect, Monod says that sci-
ence can do no more to explain the origins of life than simply to state the
obvious: Life is possible, and has originated.
The theory of information shows, in fact, that a complex form can be

"explained" only by a statement of equal or greater complexity that in-
cludes the information in the original form. Such a statement can amount
to little more than a description in some language of the original f o r m — a
mere assertion of the form's existence. Thus complex form constitutes a
barrier of ultimate frustration for those who are seeking simple explanations
of nature.
We cannot avoid this frustration by adopting the hypothesis that the in-
formation for living beings was encoded in the initial or boundary condi-
tions of the universe. This hypothesis suggests that the information for life
must have originated beyond the boundary of the universe, or at a time be-
fore the universe existed. Yet nothing can lie beyond the spatial or tem-
poral boundaries of a truly universal model, and if we envision a succession
of models of greater and greater extent, then we are stuck with an infinite
regress.
As long as we adhere to a strictly quantitative theoretical approach, we
will have to regard the information for life (or, indeed, any complex in-
formation) as simply inexplicable. One possible option, however, is to
suppose that the boundary of our model represents the point at which
mathematical description must be given up and replaced by a nonreduc-
tionistic understanding of reality. This opens up the possibility that the
origin of complex information may be understandable in a higher context
that transcends purely numerical considerations. This option is actually not
at all unreasonable a priori, but it may appear unpalatable to a person
strongly committed to the quantitative method of studying nature.
Let us therefore leave this option aside for the moment, and consider
interpretation (3). We have pointed out in Section 5.1 that the idea of
complex natural laws is not compatible with the predominant world view
of modern science. Physicists have traditionally sought to explain natural
phenomena using the simplest laws possible, and scientists in general have
tended to assume that the accepted laws of their time were final and uni-
versal. Yet we have also seen that it is very difficult to even formulate,
much less verify, an adequate representation of the laws of nature. (See
Appendix 2 and also Chapter 3.)
On closer inspection it becomes apparent that our knowledge of the laws
of nature is not final. Indeed, some eminent scientists have proposed that
the known laws may represent only a negligible fraction of the natural laws
and principles that remain to be discovered. For example, the physicist
David Bohm has observed that "the possibility is always open that there
may exist an unlimited variety of additional properties, qualities, entities,
systems, levels, etc. to which apply correspondingly new kinds of laws of

41
nature."
Now, what is really meant by a "law of nature"? We have treated such
laws as mathematical equations, and from a strictly empirical viewpoint
they are simply numerical regularities in the welter of measurable phe-
42
nomena. Yet many scientists have tended to imagine that these laws
represent a fundamental stratum of reality underlying the world of appear-
ances that confronts our senses. They have thus tended to construct a world
view in which such things as electrons and quantum fields constitute actual
reality, while such things as mountains, trees, and people are simply tran-
sient combinations of these underlying ingredients. These scientists have
been trying to understand the absolute basis of reality, and they have
hoped that this ultimate foundation would be disclosed through the study
of physical laws.
We might therefore ask, "What is the absolute reality corresponding to
the laws of nature?" We can be sure that whatever this is, it does not consist
of numbers, "bits," or mathematical operations. If the laws of nature can
be expressed in simple formulas, then we might hope to visualize this re-
ality vaguely in terms of some kind of abstract curves and surfaces. But this
is not possible if the laws of nature encompass large amounts of mathemati-
cally irreducible information specifying an elaborate hierarchy of biologi-
cal forms. Due to their very irreducibility, such laws cannot be visualized in
terms of simple mathematical constructs. From the quantitative, empirical
point of view, the reality corresponding to such laws could only be regarded
as completely enigmatic.
Purely empirical arguments cannot carry us very far beyond these rather
limited and unsatisfactory conclusions. It may be worthwhile, however, to
briefly consider some other approaches to our subject matter that people
have pursued in the past. One of the most prominent of these approaches
was widely known in Darwin's time as natural theology. According to natu-
ral theology, the presence of highly complex order in nature can be in-
terpreted as evidence of intelligent direction. The proponents of this view
depart radically from the quantitative school of thought by proposing that
the absolute foundation of nature is a primordial sentient being lying
completely outside the realm of mathematical describability.
In Darwin's time and earlier, many scientists accepted the argument that
complex design in living organisms implies the existence of a transcen-
43
dental designer. With the introduction of Darwin's theory of evolu-
tion, however, this argument was weakened. Darwin's principle of natural
selection appeared to provide an explanation of the origin of biological
form that could be based completely on simple natural laws. This explana-
tion seemed to involve fewer inexplicable elements than the hypothesis
of a primordial creator, and many people therefore came to accept it as
scientifically superior.
The analysis presented here removes this objection to natural theology.
We have seen that the origin of complex order can be explained neither by
natural selection nor by any other principle based on simple natural laws.
Natural structures and patterns of high information content are inherently
inexplicable by the reductionistic methods of quantitative analysis. The
H hypothesis that these structures have been generated by a transcendental
intelligent being is therefore in no way inferior to our alternative interpre-
tations (1), (2), and (3) as a theoretical explanation. In addition, this
hypothesis opens up the possibility that if the ultimate foundation of reality
is a sentient being, then it may be possible to acquire absolute knowledge
directly from this transcendental source.
At this point the objection might be raised that this hypothesis is not
proven by our analysis, and that, indeed, we have not rigorously proven
that biological form is characterized by high information content. If the
forms of organisms only seem complex, but actually have low informa-
tion content, then our analysis does not rule out the possibility that they
might evolve with relatively high probabilities. Thus, by adopting interpre-
tation (4), we may be able to save the idea of progressive evolution in a
comprehensible universe governed by simple natural laws.
This approach to the theory of evolution has been adopted by the infor-
mation theorist Gregory Chaitin. According to Chaitin, the fundamental
goal of theoretical biology is
to set up a nondeterministic model universe, to formally define what it means
for a region of space-time in that universe to be an organism. . . , and to
rigorously demonstrate that starting from simple initial conditions, organisms
will appear and evolve . . . in a reasonable amount of time and with high
44
probability.
This program requires living organisms to have low information content,
and Chaitin has therefore proposed that an organism is "a highly inter-
dependent region, one for which the complexity of the whole is much less
45
than the sum of the complexities of its parts."
In order to elucidate Chaitin's proposal, let us refer again to the analysis
presented in Sections 5.2 and 5.3. We showed with inequalities (18), (20),
and (22) of Section 5.3 that no pattern with an information content of more
than 101,392 (= 100,800 + 326 + 266) bits is likely to evolve in our model
system. Our upper estimate in Section 5.2 for the number of proteins in a
cell was 210,000. For the sake of argument, let us suppose that a list of
210,000 proteins has an information content of no more than 101,392 bits.
Then inequalities (12) and (13) of Section 5.2 imply that each successive
protein can be specified in terms of the preceding proteins in the list by the
46
addition of only 1.5 bits of new information on the average. In fact, (13)
and (14) imply that for at least 210,000-101,392 = 108,608 values of n, we
will have F(1 ,X _ ) = Y„. In other words, 108,608 of these proteins must be
n 1
exactly determined by the preceding proteins in the list.

Chaitin and his colleague C . H . Bennett are suggesting that relationships
of this kind indeed exist, but that they are hidden from our vision. Their
argument runs as follows: If structure has a low information content, then,
by definition, it can be specified by a short program. Yet even though a pro-
gram is short, it may run for a very long time before finishing its calcula-
47
tions. Bennett has suggested that biological structures do have very low
information content, but can be defined by short programs only if these
programs have very long running times. This means that extensive inter-
relationships must run through all biological structures, but it will not be
possible for us to demonstrate this.
If this approach is to save the idea of progressive evolution, it must be
consistently applied to the entire spectrum of biological form. A l l signifi-
cant symbolic descriptions of organic form must have a low information
content. This means that these descriptions and their subdivisions can be
systematically transformed into one another using the generating function
F. With the addition of small amounts of information, a description of, say,
the visual processes of a mammal can be converted into a description of the
mitotic machinery of cells; this in turn can be converted into a description
of abstract features of human mental ability; and so on. Yet if Bennett's
ideas are correct, it will be impossible to show these relationships by any
48
means of analysis that could conceivably be carried out by human beings.
We can conclude that the empirical method of investigation leads us into
a frustrating impasse when we try to use it to understand the nature and ori-
gin of life. Our effort to rigorously analyze the theory of evolution has led
us to consider the generating function F. It seems highly implausible that
this simple function—which is defined in a few lines in Appendix 1—can
effect the remarkable transformations we have described. There is also no
direct evidence that complex designs of organic mechanisms can be auto-
matically generated by the execution of simple computational routines.
Yet the possibility remains that this can be done by routines with such long
running times that we cannot test them.
At this point some persons might feel inclined to adopt alternative (5),
and conclude that speculation on these topics is futile. We can agree that
such a conclusion is not at all unreasonable, but we should stress that it en-
tails abandoning all confidence in the theory of evolution as an established
understanding of the origin of life. One might advocate, of course, that we
should admit our present ignorance in this area, while still maintaining the
hope that some day we will show how evolution proceeds by simple laws.
We should realize, however, that biological form may well be characterized
by high information content, as we argued in Section 5.2. If this is indeed
the case, the empirical approach will never be able to provide a simple ex-
planation of life's origins. At best, this approach will be able to produce
only a complex description of natural order. Such a description cannot be
satisfying as an explanation, and, as Bennett's ideas indicate, it can never
49
be possible to actually prove that the description is f i n a l .
These conclusions may seem to be completely negative, but we should
not end this discussion without pointing out that they also have a positive
side. If nature is indeed simple and comprehensible, then it is hard to imag-
ine any other method of studying nature than the empirical method of ex-
periment and calculation. But by showing that nature may not be so simple,
we are opening up the possibility that there may be other valid methods of
obtaining knowledge. If the absolute foundation of reality is a reservoir of
irreducible and inconceivable qualities, then the possibility arises that we
may be able to approach the absolute by methods other than dissection and
measurement. Our empirical analysis of evolutionary theory cannot prove
rigorously that this is so, and it cannot disclose how such an alternative ap-
proach could be practically realized. However, it does show that empirical
reasoning is capable of pointing beyond itself.
We have mentioned the conclusion of natural theology, that the pres-
ence of complex order in nature implies the existence of an absolute intel-
ligence. Although this conclusion is consistent with the results of our
information-theoretic analysis, it is certainly not proven by this analysis.
Nonetheless, the ideas of natural theology are valuable, for they suggest an
alternative to the empirical method of acquiring knowledge. If the absolute
basis of nature is a sentient being and we ourselves are products of this ab-
solute source, then it is possible that we may be able to relate personally
with the absolute.
In this chapter we will not discuss the ramifications of this possibility,
but we mention it simply as a concrete example of the options opened up
by our analysis. If it is to be of real interest, such an approach to absolute
knowledge must be verifiable by tangible, reproducible experience, for
otherwise it amounts to nothing more than empty speculation. Yet the
Notes 139
practical means of accomplishing this cannot be established by quantitative

analysis, and so they lie beyond the scope of the present discussion. Here
we must stop at the boundaries of the empirical realm.
Notes
1. Weinberg, " C o n c e p t u a l Foundations of the Unified Theory of Weak

and Electromagnetic Interactions," p p . 1212-1218.
2. Hawking and Israel, eds., General Relativity, p. 2 1 .
3. Slater, Quantum Theory of Molecules and Solids, Vol. 1, p. vii.
4. Watson, Molecular Biology of the Gene, p. 54.
5. Newton's Principia, p. Ixviii.
6. H e l m h o l t z , Ūber die Erhaltung der Kraft, p. 6.
7. These q u a n t u m mechanical equations are taken from von N e u m a n n ,

Mathematical Foundations of Quantum Mechanics, and Messiah,
Quantum Mechanics.
8. Fisher, The Genetical Theory of Natural Selection.
9. For example, see O p a r i n , The Origin of Life, and Orgel, The Origins
of Life.
10. Layzer, " T h e Arrow of T i m e , " p. 206.
11. This is discussed in Tolman, The Principles of Statistical Mechanics.
12. von Neumann, Theory of Self-Reproducing Automata.
13. T h e data for E. coli are taken from Watson, chap. 3.
14. Watson, p. 69.
15. Watson, p. 498.
16. Watson, p. 677.
17. Crick, "Split G e n e s and R N A Splicing," p p . 264-271.
18. Watson, p. 499.
19. Watson, p. 404.

20. T h e F i n equation (11) contains a p a r a m e t e r of l= 10,000 specifying the
lengths of the intervals Yj. In this application of F we have to change
this p a r a m e t e r to l= 16,800 bits/page.
2 1 . Watson, p. 76.
22. Satir, " H o w Cilia M o v e , " pp. 4 4 - 5 2 .
23. Staehelin and Hull, "Junctions between Living Cells," pp. 141-152.
24. Winston, "The M I T R o b o t . "
25. According to a famous conjecture of Sir A r t h u r Eddington, there are
2 5 6 79
exactly 136 x 2 1.6x 1 0 protons in the universe. (Eddington, The
Philosophy of Physical Science, p. 176.) Of course, this should not be
taken very seriously. The important point here is that l o g r w i l l be 2
fairly small, even if the total n u m b e r of atoms in the system is very

large.
26. M a c b e t h , Darwin Retried, p. 47.
27. H e r e is a brief outline or a simpler but more specialized answer. We

have estimated that the D N A coding of a higher animal will contain at
5
least 6 x 10 bits of significant information. Suppose that this coding
also contains N bits of random noise. The random noise should not
duplicate the significant information, and thus the total information
5
content of the D N A coding should be at least 6 x 10 + N. The proba-
bility that a particular animal with this coding will evolve is therefore
- N -1500,000
bounded by p = 2 10 , according to inequality (18). Now, let
P be the probability of the evolution of at least one animal with a
genome containing this fixed significant information plus N bits of
N
arbitrary noise. This P could be at most 2 times as large as p, since
N
there are only 2 possibilities for N bits of information. This means
N -1500,000
that P s 2 p 10 , which is the same bound as in inequality (20).
28. Prigogine, Nicolis, and Babloyantz, "Thermodynamics of Evolution,"

pp. 23-28.
29. Eigen and Schuster, " T h e Hypercycle," in three parts.
30. Cozzarelli, " D N A Gyrase and the Supercoiling of D N A , " pp. 953-960.
31. Watson, p p . 316-317.
32. Dixon and Webb, Enzymes, pp. 656-663.
33. Smith, "Hypercycles and the Origin of Life," pp. 445-446.
34. Yockey, "A Calculation of the Probability of Spontaneous Biogenesis
by Information T h e o r y , " pp. 377-398, and " O n the Information
Content of Cytochrome c," pp. 345-376.
Notes 141
35. Julian Huxley is one well-known proponent of the idea that evolution
is a progressive process directed by natural selection. See Huxley,
Evolution in Action.
36. Shklovskii and Sagan, Intelligent Life in the Universe.
37. Gould, "Chance Riches," p p . 3 6 - 4 4 .

38. Dobzhansky, "Darwinian Evolution and the Problem of Extra-
terrestrial Life," p. 173.
39. Simpson, This View of Life, p. 268.
40. M o n o d , Chance and Necessity, p. 43.
41. Bohm, Causality and Chance in Modern Physics, p. 133.
42. This viewpoint is expounded by the logical positivists, who hold that
one can speak meaningfully only of sense perceptions, and that all
other categories are mere verbiage. This philosophy is summarized
by the physicist Yehudah Freundlich as follows: "To us, the statement
that trains have wheels when they are not in the station (when we are
not sensing them) means that at the station they will have wheels. This
is, to us, a very satisfying solution, for having thus defined existence
we proceed to speak of wheels on the train even when it is not in the
station. In general, attributing a property to a system means that cer-
tain predictions about the system can be m a d e . " (Freundlich, "Mind,
Matter, and Physicists," p p . 130-131.) According to this doctrine, the
statement " M a n evolved from a primate ancestor" means that certain
bones may be seen in certain museums. This is far from a satisfying so-
lution to us and, we believe, to many others. Generally, scientists have
sought to acquire insight into the nature of what really exists. If em-
pirical science is indeed unable to provide such insight, we can only
conclude that its pretensions to universality are greatly overblown.
43. Gillespie, Charles Darwin and the Problem of Creation.
44. Chaitin, "Algorithmic Information T h e o r y , " p. 357.

45. Chaitin, p. 357.
46. This number refers to the average value of l{w ). By definition, w l, n n
and the choice w =0 is excluded. Therefore a w of x bits really can

n n
x
convey only log (2 -1) bits of information. This is .87 bits if x = 1 . 5 .
2
47. C o m p u t e r languages have provisions for iterating computational steps

a specified number of times. If the number of iterations in a program is

in the millions or even billions, then it may be possible to run the pro-
gram on existing computers. However, in a simple program we can
100 10^100
easily specify very large numbers of iterations, such as 1 0 or 10
It is not possible to gain practical experience with such programs, since
they would not complete their calculations in billions of years on the
fastest imaginable computer. Experience with simple programs involv-
ing moderate numbers of iterations gives us no reason to suppose that
such programs can generate detailed descriptions of life as we know it.
Should we suppose that such programs will generate descriptions of
100,000
life if their calculations are prolonged over, say, 10, iterations?
48., Mathematicians have shown that some statements can be neither

proven nor disproven, even though they have a sensible meaning and
presumably must be either true or false. These are called undecidable
statements. There are also statements, called practically undecidable,
which can be proven true, but only by proofs of such length that human
beings cannot hope to deal with them. As we might expect, complex
statements tend to be undecidable. The statement L(X) N is always
undecidable if it is true and N is large enough. If it is false, then the op-
posite statement, L(X)<N, is often practically undecidable. It is inter-
esting that we are naturally led to consider statements of this kind
when we try to rigorously formulate the theory of evolution. (For the
undecidability of L(X) N see Chaitin, "Algorithmic Information
Theory," pp. 350-359. A general account of undecidability is given in
Jones, "Recursive Undecidability—An Exposition," pp. 724-738.)
49. This would require a formal proof of L(X) N. See note 48.
Chapter 6
Chance and
The Unity of Nature
Throughout human history, philosophers and seekers of knowledge
have sought to discover a single fundamental cause underlying all the
phenomena of the universe. Since the rise of Western science in the late
Renaissance, many scientists have also felt impelled to seek this ultimate
goal, and they have approached it from their own characteristic perspec-
tive. Western science is based on the assumption that the universe can be
understood mechanistically—that is, in terms of numbers and mathema-
tical formulas—and Western scientists have therefore searched for an
ultimate, unified mathematical description of nature.
This search has gone through many vicissitudes, and many times scien-
tists have felt that a final, unified theory was nearly within their grasp. Thus
in the beginning of the nineteenth century Pierre Simon de Laplace could
contemplate Newton's laws and declare that " A l l the effects of nature are
only the mathematical consequences of a small number of immutable
1
laws." By the turn of the century many new concepts and discoveries had
been incorporated into the science of physics, and Laplace's simple picture
of the laws of nature had become superannuated. At about this time, how-
ever, Albert Einstein embarked on a much more sophisticated and ambi-
tious program of unification. His goal was to explain all the phenomena of
the universe as oscillations in one fundamental "unified field."
But even while Einstein was working on this project, revolutionary de-
velopments in the science of physics were rendering his basic approach ob-
solete. For several decades a bewildering welter of new discoveries made
the prospect of finding an ultimate theory seem more and more remote.
But the effort to find a unified theory of nature has continued, and in 1979
three physicists (Sheldon Glashow, Abdus Salam, and Steven Weinberg)
won the Nobel Prize in physics for their effort in partially tying together
some of the disparate elements of current physical theories. On the basis of
their work, many scientists are now optimistically anticipating the develop-
ment of a theory that can explain the entire universe in terms of a single
quantum field governed by a universal force law.
The scientists' search for a unified explanation of natural phenomena
begins with two main hypotheses. The first of these is that all the diverse
143
144 Chance and the Unity of Nature
phenomena of nature derive in a harmonious way from some ultimate, uni-

fied source. The second is that nature can be fully explained in terms of
numbers and mathematical laws. As we have pointed out, the second of
these hypotheses constitutes the fundamental methodological assumption
of modern science, whereas the first has a much broader philosophical
character.
Superficially, these two hypotheses seem to fit together nicely. A simple
system of equations appears much more harmonious and unified than a
highly complicated system containing many arbitrary, unrelated expres-
sions. So the hypothesis that nature is fundamentally harmonious seems to
guarantee that the ultimate mathematical laws of nature must be simple
and comprehensible. Consequently, the conviction that nature possesses
an underlying unity has assured many scientists that their program of
mechanistic explanation is feasible.
We will show in this chapter, however, that these two hypotheses about
nature are actually not compatible. To understand why this is so, we must
consider a third feature of modern scientific theories—the concept of
chance.
As we carefully examine the role chance plays in mechanistic explana-
tions of nature, we shall see that a mechanistic theory of the universe must
be either drastically incomplete or extremely incoherent and disunited. In
mechanistic models of reality there is, in fact, a fundamental conflict be-
tween unity and diversity, and many scientists have attempted to resolve
this conflict by asserting that diversity arises by chance. Unfortunately,
however, the concept of chance as an active agency is fallacious. Once
the fallacy is exposed, it becomes clear that we must give up either the
goal of mechanistically explaining the universe, or the idea that there is an
essential unity behind the phenomena of nature.
At the end of this chapter we will examine the conflict between unity and
diversity from a broader philosophical perspective. We observe that even
though this conflict cannot be resolved by a mechanistic theory, we are
already familiar with a natural phenomenon that simultaneously exhibits
the features of both unity and diversity. This is the phenomenon of con-
scious awareness. We go on to show how the Vedic literature presents a
unified nonmechanistic explanation of the phenomenal universe based on
the concept of universal consciousness.
6.1 Statistical Laws and Their Role in Modern Physics

In this section we will examine how the concept of chance is employed in
mechanistic theories. Such theories are normally formulated in the mathe-
Statistical Laws and Their Role in Modern Physics 145
matical language of physics, and they involve many complicated technical

details. Yet the basic concepts of chance and natural law in current physical
theories readily lend themselves to illustration by simple examples. We will
therefore briefly contemplate a few such examples and then draw some
general conclusions about universal mechanistic theories.
Since we are interested in universal theories, we will want to consider
how the concept of chance is to be understood in the setting of the universe
as a whole. We will therefore introduce a simple model "universe" that can
be used for posing questions about the nature of chance. Figure 1 depicts
this model. The model consists of a box with a window that always displays
either a figure 0 or a figure 1. The nature of this box is that during each con-
secutive second the figure in the window may either remain unchanged or
else change exactly once at the beginning of that second. We can thus de-
scribe the history of the model universe by a string of zeros and ones repre-
senting the successive figures appearing in the window during successive
seconds. Figure 2 depicts a sample history.
Figure 1. This device displays a figure of zero or one that can change from second to
second. We shall regard it as a model "universe" and use it to illustrate the concepts
of random events and universal statistical laws.
Figure 2. A history of the model universe spanning a total of 979 seconds.
Let us begin by considering how the concept of chance could apply to this
model universe. For example, suppose we are told that the model universe
obeys the following statistical law:
The zeros and ones appear randomly in the window, independently of one
another. During any given second the probability is 50% that the window will
display a one and 50% that it will display a zero.
How are we to interpret this statement? As we shall see, its interpretation

involves two basic questions: the practical question of how we can judge
whether or not the statement is true, and the broader question of what the
statement implies about the nature of the model universe.
The answer to the first question is fairly straightforward. We would say
that the statement is true of a particular history of ones and zeros if that
history satisfied certain statistical criteria. For example, if the probability
for the appearance of one is to be 50%, we would expect roughly 50% of
the figures in the historical sequence to be ones. This is true of the sample
history in Figure 2, where the percentage of ones is 49.4%.
We could not, however, require the percentage of ones to be exactly
50%. If the sequence itself is random, the percentage of ones in the se-
quence must also be random, and so we would not expect it to take on some
exact value. But if the percentage of ones were substantially different from
5 0 % , we could not agree that these ones were appearing in the window
with a probability of 50%.
In practice it would never be possible for a statistical analyst to say defi-
nitely that a given history does or does not satisfy our statistical law. A l l he
could do would be to determine a degree of confidence in the truth or fal-
sity of the law as it applied to a particular sequence of ones and zeros. For
example, our sample history is 979 digits long. For a sequence of this length
to satisfy the law, we would expect the percentage of ones to fall between
46.8% and 53.2%. (These are the "95% confidence limits.") If the percen-
tage did not fall within these limits, we could take this failure as an indica-
tion that the sequence did not satisfy the law, but we could not assert this as
a definite conclusion.
We have seen that the sample history consists of approximately 50%
ones. This observation agrees with the hypothesis that this sequence satis-
fies our statistical law, but it is not sufficient to establish this, for there are
other criteria such a sequence must meet. In general, if a sequence is to be
considered random or disorderly, we would not expect any particular pat-
tern to appear within the sequence with unusual prominence. This means
that for each positive number n, all of the 2" subsequences of length n
should appear in our sequence with nearly equal frequencies.
For example, if the sample history were indeed a random sequence, we
would expect each of the subsequences 00, 0 1 , 10, and 11 to appear with a
frequency of roughly 25%. In fact, the frequencies of these subsequences
are 25.6%, 24.7%, 25.4%, and 24.3% respectively, and these frequencies
conform with the hypothesis that our history satisfies the statistical law. As
before, we cannot expect the frequency to take on an exact value. We can
at most determine a degree of confidence in our statistical hypothesis by
measuring how closely the observed frequencies of various subsequences
2
match their expected values.
So in practical terms we can interpret our statistical law as an approxi-
mate statement about the relative frequency of various patterns of ones
and zeros within a larger sequence of ones and zeros. If statistical laws were
never attributed a deeper meaning than this, the concepts of randomness
and statistical law might seem of little interest. However, because of an ad-
ditional interpretation commonly given them, these concepts are actually
of great significance in modern science, and particularly the science of
physics. This interpretation becomes clear in the following reformulation
of our statistical law, as understood from the viewpoint of modern physics:
The box contains some apparatus that operates according to definite laws of
cause and effect and that determines which figures will appear in the window.
But in addition to its predictable, causal behavior, this apparatus periodically

undergoes changes that have no cause and that cannot be predicted, even in
principle. The presence of a one or a zero in the window during any given sec-
ond is due to an inherently unpredictable, causeless event. Yet it is also true
that ones and zeros are equally likely to appear, and thus we say that their
probability of appearance is 50%.
In this formulation, our statistical law is no longer simply a statement about

patterns of ones and zeros in a sequence. Rather, it now becomes an asser-
tion about an active process occurring in nature—a process that involves
absolutely causeless events. Such an unpredictable process is said to be a
"random process" or a process of "absolute chance."
We can bring out the implications of this interpretation of chance by con-
sidering the following statement from a standard textbook of probability
theory: "The fact that in a number of instances the relative frequency of
random events in a large number of trials is almost constant compels us to
presume the existence of certain laws, independent of the experimenter,
3
that govern the course of these phenomena." The author of this text is
arguing that because a sequence exhibits certain regularities, we can con-
clude that it obeys a statistical law. If we interpret statistical laws as state-
ments about the distribution of ones and zeros in a sequence, then this
conclusion is certainly justifiable. But if we view "random events" as inher-
ently causeless and unpredictable natural phenomena, we then find our-
selves in the peculiar position of regarding lawlike regularities as evidence
for the occurrence of events that, by definition, obey no law at all.
At first glance this second interpretation of the concept of randomness
may seem quite strange, even self-contradictory. Nonetheless, since the
development of quantum mechanics in the early decades of the twentieth
century, this interpretation has occupied a central place in the modern
scientific picture of nature. According to quantum mechanics, almost all
natural phenomena involve "quantum jumps" that occur by absolute, or
causeless, chance. At present many scientists regard the quantum theory as
the fundamental basis for all explanations of natural phenomena. Con-
sequently, the concept of absolute chance is now an integral part of the
scientific world view.
The role absolute chance plays in the quantum theory can be illustrated
by the classical example of radioactive decay. Let us suppose the model
universe contains some radioactive atoms, a Geiger counter tube, and
some appropriate electrical apparatus. As the atoms decay they trigger the
Geiger counter and thereby influence the apparatus, which in turn controls
the sequence of figures appearing in the window. We could arrange the
apparatus so that during any given second a one would appear in the win-
dow if a radioactive decay occurred at the beginning of that second, and

otherwise a zero would appear. By adjusting the amount of radioactive
substance, we could control the average rate at which the counter was
triggered and thus assure that the figure one would appear approximately
50% of the time.
If the apparatus were adjusted in this way, we would expect from obser-
vational experience that the sequence of ones and zeros generated by the
model universe would satisfy our simple statistical law. Modern physicists
interpret this predictable statistical behavior as evidence of an underlying
process of causeless chance. Although they analyze the operation of the ap-
paratus in terms of cause and effect, they regard the decay of the atoms
themselves as fundamentally causeless, and the exact time at which any
given atom decays as inherently unpredictable. This unpredictability im-
plies that the sequence of ones and zeros generated by the model should
follow no predictable pattern. Thus the hypothesis of causeless chance
provides an explanation of the model's statistical behavior.
If we analyze the above example of a physical system, we can see that
it involves a mixture of two basic elements: determinism and absolute
chance. In this example we assumed that the electrical apparatus followed
deterministic laws, whereas we attributed the decay of the radioactive
atoms to absolute chance. In general, the theories of modern physics entail
a combination of these two elements. The deterministic part of the theory
is represented by mathematical equations describing causal interactions,
and the element of chance is represented by statistical laws expressed in
terms of probabilities.
When some scientists view natural phenomena in the actual universe as
obeying such combined deterministic and statistical laws, they show a
strong tendency to suppose that the phenomena are governed by these
laws, and by nothing else. They are tempted to imagine that the laws cor-
respond directly to a real underlying agency that produces the phenomena.
Once they visualize such an agency, they naturally think of it as the en-
during substantial cause, and they regard the phenomena themselves as
ephemeral, insubstantial effects.
Thus the physicist Steven Weinberg refers to the theories of physics as
"mathematical models of the universe to which at least the physicists give a
4
higher degree of reality than they accord the ordinary world of sensation."
Following this line of thinking, some researchers are tempted to visualize
ultimate mathematical laws that apply to all the phenomena of the universe
and that represent the underlying basis of reality. Many scientists regard
the discovery of such laws as the final goal of their quest to understand
nature.
Up until now, of course, no one has formulated a mathematically con-

sistent universal theory of this kind, and the partial attempts that have
been made involve a formidable tangle of unresolved technical difficul-
ties. Nonetheless, even the most sophisticated theories of modern phys-
ics entail statistical laws, and are firmly based on the concept of absolute
chance. As we shall see, this simple fact constitutes a fatal flaw in the physi-
cists' program of achieving a complete and unified understanding of na-
ture. The concept of absolute chance is fundamentally fallacious, and any
theory based on it must also be fallacious, regardless of how technically
sophisticated it may be.
6.2 The Illusion of Absolute Chance

We can uncover the fallacy in the concept of absolute chance by examin-
ing the model universe more closely. If we do this we find that the laws of
this model adequately describe the deterministic functioning of the electri-
cal apparatus, and also the statistical properties of the sequence of radio-
active decays. They do not, however, have anything to say about the details
of this sequence. Of course, according to the theory of the model, this
sequence cannot be expected to display any meaningful patterns, and thus
it might seem pointless to inquire about these details. Nonetheless, for
the sake of clarity we should consider the following questions: Can our
theoretical description of the model universe really be considered com-
plete or universal? In order for the description to be complete, wouldn't
we need to incorporate into it a detailed description of the actual sequence
of radioactive decays?
Considering these questions, we note that if we were to augment the
theory in this way, then by no means could we consider the resulting en-
larged theory unified. It would consist of a short list of basic laws followed
by a very long list of data displaying no coherent pattern. Yet, if we did not
include the exact sequence of radioactive decays, we would have to admit
that the theory was incomplete in that it failed to account for this detailed
information. Clearly, we could consider such a theory complete only if
we adopted a standard of completeness that enabled us to ignore most of
the detailed features of the very phenomena the theory was intended to
describe.
Now, when we consider the concept of absolute chance, we see that it
seems to provide just such a standard of completeness. The idea that a se-
quence of events is generated by causeless chance seems intuitively to
imply that these events should be disorderly, chaotic, and meaningless. We
would not expect one totally random sequence to be distinguishable in any
The Illusion of Absolute Chance 151
significant way from any of the innumerable other random sequences

having the same basic statistical properties. We would expect the details of
the sequence to amount to nothing more than a display of pointless
arbitrariness.
This leads naturally to the idea that one may consider a theoretical
description of phenomena complete as long as it thoroughly accounts for
the statistical properties of the phenomena. According to this idea, if
the phenomena involve random sequences of events—in other words,
sequences satisfying the statistical criteria for randomness—then these se-
quences must be products of causeless chance. As such, their detailed
patterns must be meaningless and insignificant, and we can disregard
them. Only the overall statistical features of the phenomena are worthy of
theoretical description.
This method of defining the completeness of theories might seem satis-
factory when applied to the example of radioactive decay. Certainly the
observed patterns of atomic breakdown in radioactive substances seem
completely chaotic. But let us look again at the sample history of the model
universe depicted in Figure 2. As we have already indicated, this history
satisfies many of the criteria for a random sequence that can be deduced
from our simple statistical law. It also appears chaotic and disorderly. Yet if
we examine it more closely, we find that it is actually a message expressed
in binary code.
When we decipher this message it turns out, strangely enough, to consist
of the following statement in English:
The probability of repetition of terrestrial evolution is zero. The same holds
for the possibility that if most life on earth were destroyed, the evolution
would start anew from some few primitive survivors. That evolution would be
most unlikely to give rise to new man-like beings.
What are we to make of this? Could it be that by some extremely improba-

ble accident, the random process corresponding to our simple statistical
law just happened to generate this particular sequence? Using the law,
we can calculate the probability of this, and we obtain a percentage of
.000 . . . (292 zeros) . . . 0001.
The answer, of course, is that we did not actually produce the sequence
in Figure 2 by a random process. That a sequence of events obeys a sta-
tistical law does not imply that a process of chance governed by this law
actually produced the sequence. In fact, the sequence in Figure 2 demon-
strates that at least in some situations, the presence of a high degree of
randomness in a sequence calls for a completely different interpretation.
When we consider the method used to construct this sequence, we see that
its apparent randomness results directly from the fact that it encodes a large
amount of meaningful information.
We produced the sequence in Figure 2 by a technique from the field of
communications engineering known as "data compression." In this field,
engineers confront the problem of how to send as many messages as pos-
sible across a limited communications channel, such as a telephone line.
They have therefore sought methods of encoding messages as sequences of
symbols that are as short as possible but can still be readily decoded to
reproduce the original message.
In 1948 Claude Shannon established some of the fundamental principles
5
of communications engineering. He showed that each message has a cer-
tain information content, which can be expressed as a number of "bits," or
binary ones and zeros. If a message contains N bits of information, we can
encode it as a sequence of N or more ones and zeros, but we cannot encode
it as a shorter sequence without losing part of the message. When we en-
code the message as a sequence of almost exactly N ones and zeros, its
density of information is maximal, and each zero or one carries essential
information.
Shannon showed that when encoded in the shortest possible sequence, a
message appears to be completely random. The basic reason for this is that
if patterns of ones and zeros are to be used in the most efficient possible
way to encode information, all possible patterns must be used with roughly
equal frequency. Thus the criteria for maximal information density and
maximal randomness turn out to be the same.
Figures 3 and 4 show the effects of information compression for the mes-
sage encoded in Figure 2. Figure 3 illustrates some of the characteristics of
an uncompressed binary encoding of this message. The bar graph in this
figure represents the frequency distribution for five-bit subsequences, each
representing a letter of the English text. This distribution clearly does not
follow the bell-shaped curve we would expect for a random sequence.
However, when we encode the message in compressed f o r m , as in Figure 2,
we obtain the distribution shown in Figure 4. Here we see that simply by
encoding the message in a more succinct form, we have greatly increased its
6
apparent randomness.
We can conclude that it is not justifiable to insist upon absolute chance as
an explanation of apparent randomness in nature. If a sequence of events
exhibits the statistical properties of randomness, this may simply mean that
it contains a large amount of significant information. Also, if a sequence
exhibits a combination of random features and systematic features, as with
our text before compression, this may reflect the presence of significant in-
Chance and Evolution 153
formation in a less concentrated form. In either case, we would clearly be

mistaken to disregard the details of the sequence, thinking them simply
products of chance.
In summary, we recall that a sequence of events is generally called ran-
dom if it satisfies the statistical criterion that all possible subsequences
should appear with roughly equal frequency. Many scientists have inter-
preted the occurrence of random sequences in nature as evidence for the
existence of a natural process of causeless chance. Yet we have observed
that the criterion for maximal information content in a sequence is the
same as the criterion for randomness, and a sequence containing a high
density of meaningful information certainly cannot be attributed to abso-
lute chance. Since the idea of inferring lawlessness from lawful behavior
seems contradictory anyway, we can only conclude that the concept of ab-
solute chance is fallacious. This concept simply serves to give us an illusion
of understanding when we deal with phenomena that we can describe but
cannot explain.
6.3 Chance and Evolution
At this point let us consider how these observations bear on the actual
universe in which we live. Could it be that while focusing on ultimate
mechanistic laws, modern scientists are disregarding some significant in-
formation encoded in the phenomena of nature? In fact, this is the implica-
tion of the sequence in Figure 2 when we decode it and perceive its higher
meaning—namely, as a statement about human evolution. The source
7
of this statement is the prominent evolutionist Theodosius Dobzhansky,
who here expresses a view held widely among researchers in the life sci-
ences. Dobzhansky is visualizing the origin of human life in the context of
an underlying physical theory that involves combined processes of causa-
tion and chance. He is expressing the conviction that although such pro-
cesses have generated the highly complex forms of human life we know,
they nonetheless have a zero probability of doing so.
No one has shown, of course, that the universe as a whole, or even the
small part of it we inhabit, really does obey some fundamental mechanis-
tic laws. Yet suppose, for the sake of argument, that it does. In effect,
Dobzhansky is asserting that from the viewpoint of this ultimate universal
theory, the detailed information specifying the nature and history of
8
human life is simply random noise. The theory will be able to describe only
broad statistical features of this information, and will have to dismiss its
essential content as the vagaries of causeless chance.
/
The underlying basis for Dobzhansky's conviction is that he and his

fellow evolutionists have not been able to discern in nature any clearly de-
finable pattern of cause and effect that enables them to deduce the forms of
living organisms from basic physical principles. Of course, evolutionists
customarily postulate that certain physical processes called mutation and
natural selection have produced all living organisms. But their analysis of
these processes has given them no insight into why one form is produced
and not some other, and they have generally concluded that the appear-
ance of specific forms like tigers, horses, and human beings is simply a
matter of chance. This is the conclusion shown, for example, by Charles
Darwin's remark that "There seems to be no more design in the variability
of organic beings, and in the action of natural selection, than in the course
9
which the wind blows."
Now, one might propose that in the future it will become possible to de-
duce the appearance of particular life forms from basic physical laws. But
how much detail can we hope to obtain from such deductions? Will it be
possible to deduce the complex features of human personality from funda-
mental physical laws? Will it be possible to deduce from such laws the de-
tailed life histories of individual persons and to specify, for example, the
writings of Theodosius Dobzhansky? Clearly there must be some limit to
what we can expect from deductions based on a fixed set of relatively
simple laws.
These questions pose a considerable dilemma for those scientists who
would like to formulate a complete and unified mechanistic theory of the
universe. If we reject the unjustifiable concept of absolute chance, we see
that a theory—if it is to be considered complete—must directly account for
the unlimited diversity actually existing in reality. Either scientists must be
satisfied with an incomplete theory that says nothing about the detailed
features of many aspects of the universe—including life—or they must be
willing to append to their theory a seemingly arbitrary list of data that
describes these features at the cost of destroying the theory's unity.
We can further understand this dilemma by briefly considering the physi-
cal theories studied before the advent of quantum mechanics and the for-
mal introduction of absolute chance into science. Based solely on causal
interactions, these theories employed the idea of chance only to describe
an observer's incomplete knowledge of the precisely determined flow of
actual events. Although newer developments have superseded these theo-
ries, one might still wonder how effective they might be in providing a uni-
fied description of nature. We shall show by a simple example that these
theories are confronted by the same dilemma that faces universal theories
based on statistical laws.
Chance and Evolution 157
Figure 5 depicts a rectangular array of evenly spaced spheres. Let us sup-

pose that the positions of these spheres are fixed and that the array extends
in all directions without limit. We shall consider the behavior of a single
sphere that moves according to the laws of classical physics and rebounds
elastically off the other spheres. We can imagine that once we set the single
sphere into motion, it will continue to follow a zigzag path through the
fixed array of spheres.
Figure 5 illustrates how a slight variation in the direction of the moving
sphere can be greatly magnified when it bounces against one of the fixed
spheres. On successive bounces this variation will increase more and more,
Figure 5. An example illustrating a simple deterministic theory. The moving sphere

bounces elastically against fixed spheres in an infinite rectangular array. A slight
variation in the direction of the moving sphere is quickly amplified into a large
change in the sphere's path.
and we would therefore have to know the sphere's initial direction of mo-
tion with great accuracy to predict its path correctly for any length of time.
For example, suppose the moving sphere is going sixty miles per hour, and
the dimensions of the spheres are as shown in the figure. To predict the
moving sphere's path from bounce to bounce for one hour, we would have
to know its original direction of motion (in degrees) with an accuracy of
10
roughly two million decimal places. We can estimate that a number with
this many decimal digits would take a full 714 pages to write down.
In effect, the number representing the initial direction of the sphere con-
stitutes a script specifying in advance the detailed movements of the sphere
for one hour. To specify the sphere's movements for one year, this script
would have to be expanded to more than six million pages. We can there-
fore see that this simple deterministic theory can provide complete predic-
tions about the phenomena being studied—namely, the movements of the
sphere—only if a detailed description of what will actually happen is first
built into the theory.
We can generalize the example of the bouncing sphere by allowing all the
spheres to move simultaneously and to interact not merely by elastic colli-
sion but by force laws of various kinds. By doing this we obtain the classical
Newtonian theory of nature mentioned by Pierre Simon de Laplace in the
quotation cited at the beginning of this chapter. Laplace and many other
scientists of his time wished to account for all phenomena by this theory,
which was based entirely on simple laws of attraction and repulsion
between material particles.
Let us therefore consider what this theory implies about the origin of life.
Although it is more complicated than our simple example, this theory has
some of the same characteristics. To account for life as we know it, the
theory would have to incorporate billions of numbers describing the state
of the world at some earlier time, and the entire history of living beings
would have to be encoded in the higher-order decimal digits of those
11
numbers. Some of these decimal digits would encode the blueprints for a
future rhinoceros, and others would encode the life history of a particular
human being.
These digits would encode the facts of universal history in an extremely
complicated way, and as far as the theory is concerned this encoded infor-
mation would be completely arbitrary. This might tempt an adherent of
the theory to abandon the idea of strict determinism and say—perhaps
covertly—that the encoded information must have arisen by absolute
chance (see note 8). Yet we have seen that this is a misleading idea, and it
certainly has no place in a theory based solely on causal interactions. A l l we
can realistically say in the context of this theory is that the facts of universal
The Paradox of Unity and Diversity 159
history simply are what they are. The theory can describe them only if a
detailed script is initially added to it.
We can conclude that the prospects for a simple, universal mechanistic
theory are not good. Once we eliminate the unsound and misleading idea
of absolute chance, we are confronted with the problem of accounting for
an almost unlimited amount of detailed information with a finite system of
formulas. Some of this information may seem meaningless and chaotic, but
a substantial part of it is involved with the phenomena of life, and this part
includes the life histories of all scientific theorists. We must regard a theory
that neglects most of this information as only a partial description of some
features of the universe. Conversely, a theory that takes large amounts of
this information into account must be filled with elaborate detail, and it can
hardly be considered simple or unified.
6.4 The Paradox of Unity and Diversity

It seems that on the platform of finite mathematical description, the
ideal of unity is incompatible with the diversity of the real world. To ac-
count for diversity, a mechanistic theory has to contain a large amount of
irreducible, detailed information. This is certainly true of diversity that we
might consider unimportant, such as the particular pattern of waves on the
surface of the ocean, or the swirling patterns of leaves blowing in the wind.
It is also true of patterns of diversity to which we tend to attach more sig-
nificance, such as the complex bodily designs of living organisms, or the
productions of advanced human culture.
There is always the hope that one may be able to reduce the apparent
complexity of a pattern by discovering a simple underlying law that can be
used to compute the pattern. The discovery of such laws has been the goal
of physics since the days of Galileo and Newton, and such discoveries can
be viewed as instances of data compression, in which large masses of raw
data are represented by codes consisting of a few formulas and some initial
and boundary conditions. Yet we have seen in this chapter and in Chapter 5
that we cannot expect the total information describing the real world to
be unlimitedly compressible. We have every reason to suppose that the
information describing complex phenomena—such as the phenomena of
life—can be compressed only to a certain degree, leaving a large residue of
information that cannot be represented in simpler terms.
If we examine the concepts of unity and diversity, we can see that in
modern science unity has figured as an underlying philosophical goal,
whereas diversity is an undeniable, observable feature of the world. A n d so
we might conclude that the goal of finding unity is a mere chimera—and
simply resign ourselves to accepting that the world contains a large amount
of incomprehensible complexity that can be described but not explained.
Unfortunately, this conclusion would bar us from ever understanding the
most important features of life, and of human life in particular.
Here we shall set aside the mechanistic paradigm of modern science and
pursue a different approach to the problem of finding a unified cause un-
derlying the phenomena of nature. Our starting point is the observation
that we are already familiar with a phenomenon that simultaneously
exhibits the features of both unity and diversity. This is the phenomenon of
conscious awareness.
Consider what happens when we look out through a window and observe
a distant scene. We are simultaneously aware of distinct elements of the
scene—such as buildings, trees, and clouds—and we are also aware of
our accompanying thoughts and feelings. It is true that our awareness of
the details of the scene may be quite imperfect. (Certainly our memory is
imperfect, for if we are asked to describe the scene after seeing it, we may
be able to give only a very incomplete and distorted account.) Nonetheless,
it must be admitted that while we are observing the scene, our conscious
awareness does simultaneously encompass many separate features.
It may be argued that the eye scans quickly from one part of the scene to
another, and that the apparent unity of the picture we perceive is simply
due to the rapid blending together of numerous distinct picture elements.
However, it is just as hard to account for the unified conscious perception
of events in a temporal sequence as it is to account for the simultaneous
perception of different parts of a picture.
A photograph of the scene consists of many distinct colored grains that
are not tied together in a unified way. The sequence of electrical pulses
emerging from a television camera may represent the scene as a temporal
pattern, but these pulses are also not unified. If the information represent-
ing the scene is processed by a computer, it can be transformed within the
computer's memory into a variety of different forms, and these can succeed
one another at a rapid rate. Yet neither these representations nor their
12
pattern of succession in time will ever exhibit real u n i t y . As we have
observed, information can be compressed only to a certain degree, and the
residue of incompressible information represents intractable diversity that
cannot be treated mechanistically in a unified way.
It is evident that our conscious awareness possesses features of simul-
taneous unity and diversity that are paradoxical from the mechanistic
point of view. We have already observed in Part I that consciousness is a
feature of reality that cannot be explained in mechanistic terms. There we
introduced a nonmechanistic understanding of consciousness based on

the Bhagavad-gītā and other texts from the Vedic literature of India.
Here we would like to introduce another element of the world view of the
Bhagavad-gītā. As we shall see, the Bhagavad-gītā does provide a descrip
tion of how the variegated phenomena of the universe arise from a single
unified source. This is done by introducing the nonmechanistic concept of
universal consciousness.
According to the Bhagavad-gītā, conscious personality is the irreducible
basis of reality. The ultimate source of all phenomena is understood to be a
single Supreme Personality, who is known by many personal names such as
Krsna, Mādhava, and Hrsīkeśa. This primordial person is fully endowed
with consciousness, senses, knowledge, w i l l , and all other personal fea
tures. According to the Bhagavad-gītā, all these attributes are absolute and
transcendental to matter. It is not possible to reduce them to the mathe
matically describable interaction of some simple entities that can be repre
sented by sets of numbers. Rather, all the variegated phenomena of the
universe, including the phenomenon of life, are manifestations of the
energy of the Supreme Person, and they can be fully understood only in
relation to this original source.
We should stress immediately that the Supreme Person, as portrayed in
the Bhagavad-gītā, can be neither investigated nor adequately understood
by empirical research based on ordinary sense data and conducted within
the framework of mechanistic thought. We do not maintain that the empiri
cal arguments presented in this chapter and in Chapter 5 prove the exis
tence of such a transcendental being. Rather, these arguments disclose
some insurmountable barriers blocking the path of conventional scientific
investigation, and they give some hints as to what may lie beyond.
These arguments show that the forms of living beings and the patterns of
human culture are inexplicable from the mechanistic point of view. They
also suggest that the information for these highly complex forms and pat
terns is just as much a part of reality as the information for the simple natu
ral laws studied by physicists. Now, if these complex patterns are indeed
manifestations of a single, unified whole, what can we say about the nature
of that whole? Such a mechanistically impossible entity would constitute an
absolute, unified source of irreducible information pertaining to life and
personality. This at least suggests that the world has been generated by a
supreme intelligent being.
It is therefore interesting to consider the idea that a universal conscious
being may constitute the unified foundation of reality. To elucidate this
idea, we shall briefly discuss the characteristics of the Supreme Person, as
understood in the Vedic literature. We shall begin by considering the first

śloka of the īśopanisad.
om pūrnam adah pūrnam idam

pūrnāt pūrnam udacyate
pūrnasya pūrnam ādāya
pūrnam evāvaśisyate
"The Personality of Godhead is perfect and complete, and because He is
completely perfect, all emanations from H i m , such as this phenomenal
world, are perfectly equipped as complete wholes. Whatever is produced
of the complete whole is also complete in itself. Because He is the com
plete whole, even though so many complete units emanate from H i m , He
remains the complete balance." 1 3
This conception of completeness is reminiscent of certain ideas in the
mathematical theory of sets, wherein we find that one infinite set can be ,
taken away from another infinite set without depleting it. For example y j
if we take the set of all odd integers away from the set of all integerśfwe
obtain the set of even integers. If these are then divided by two, we are left
once again with the set of all integers.
The theory of infinite sets can also be used to illustrate, in an approxi
mate way, the idea of an entity that is simultaneously differentiated and
unified. We can regard a finite set of, say, one hundred points as essentially
disunited, since any part of it has fewer points than the whole and is there
fore different from the whole. In this sense, the only unified finite set is the
set consisting of exactly one point. In contrast to this, consider a continuous
line one unit in length. If we select any small segment of this line, no matter
how short, then the entire line can be obtained by expanding this segment.
Thus, the line has unity in the sense that it is equivalent to its parts. If we
consider the reason for this, we can see that it depends on the fact that the
line has infinitely many parts.
Although this is a crude example, it will serve as a metaphor to illustrate
the difference between the Supreme Person and the hypothetical physical
processes depicted in mechanistic theories. A mechanistic theory based on
a finite system of mathematical expressions can be truly unified only if it
can be reduced to one symbol—a single binary digit of " 0 " or " 1 . " Of
course, a theory that described the world in this way would say nothing
at all, and theorists have had to settle for the goal of seeking the simplest
possible theory that can adequately describe nature. Unfortunately, we
have seen that the simplest adequate theory must be almost unlimitedly
complex.
It follows that the fundamental basis of nature could possess no unity if it

truly could be described by a mechanistic theory. In contrast, the Supreme
Person is understood in the Vedic literature to be a complete unit, for even
though He possesses unlimited variegated features, He is nondifferent
from His parts. This is indicated by the following description of Krsna from
the Brahma-samhitā.
eko 'py asau racayitum jagad-anda-kotim

yac-chaktir asti jagad-anda-cayā yad-antah
andāntara-stha-paramānu-cayāntara-stham
govindam ādi-purusam tarn aham bhajāmi
"He is an undifferentiated entity, in that there is no distinction between the

potency and the'possessor thereof. In His work of creation of millions of
worlds, His potency remains inseparable. A l l the universes exist in H i m ,
and He is present in His fullness in every one of the atoms that are scattered
through the universe, at one and the same time. Such is the primeval Lord
whom I adore." 1 4
These characteristics of Krsna are dimly reflected in our example of the
line, but there is a significant difference. The equivalence of the line to its
parts depends on an externally supplied operation of expansion, and thus
the unity of the line exists only in the mind of an observer (as does the line
itself, for it is only an abstraction). In contrast, the identity of Krsna with
His parts is inherent in the reality of Krsna Himself, and therefore His unity
is natural and complete. We cannot, of course, conceive of this in terms of
mathematical examples, but we may obtain some inkling of what it means
by considering the idea of universal consciousness.
We should emphasize the distinction between universal and individual
consciousness. It is stated in the Vedic literature that just as a drop of salt
water is qualitatively the same as the ocean, so the consciousness of the in
dividual person is qualitatively the same as the universal consciousness of
Krsna. However, there must be a tremendous quantitative difference,
for what person can honestly say that he is conscious of the entire universe
and whatever lies beyond?
Nonetheless, some philosophers and scientists have declared that indi
vidual and universal consciousness are one and the same. One notable sci
entist who has come to this conclusion is the physicist Erwin Schr dinger,
who said that " I — I in the widest meaning of the word . . . — a m the per
son, if any, who controls the 'motion of the atoms' according to the Laws of
Nature." 1 5 Schr dinger believed that matter behaves strictly in accordance
with the "Laws of Nature," and he reconciled this with his perception of
himself as an active conscious being by identifying himself with the uni

versal consciousness underlying all phenomena. He supported his asser
tion by saying that according to the Upanisads, "the personal self equals the
omnipresent, all-comprehending eternal self." 1 6 We note, however, that
while the Upanisads and other Vedic literatures do assert the qualitative
equality of the supreme conscious being and the individual conscious
selves, they do not maintain that they are identical. 1 7
Since Schr dinger and others have found the relationship between the
Supreme Person and the laws of nature to be a source of perplexity, let us
consider this relationship in greater detail. As we have indicated, the phi
losophy of the Bhagavad-gītā holds that the final cause underlying all phe
nomena is the Supreme Person, Krsna. According to the Bhagavad-gītā,
the matter and energy studied in modern physics originally emanate from
Krsna, and their actions and transformations are completely determined
by His will. A l l phenomena are under direct conscious control, but since
the consciousness of Krsna is unlimited, only an infinitesimal fraction of
His attention is required to run the material universe.
At first glance, this might seem to be incompatible with our knowledge
of physics. Some phenomena of nature do appear to follow rigid, deter
ministic laws, and this might seem to rule out the possibility that they are
under direct conscious control. However, there is no real contradiction
here. From a mechanistic viewpoint, the actions of an intelligent being
must appear as patterns in the phenomena of nature. Just as a human being
can create architectural designs exhibiting simple patterns, so the Supreme
Person can easily impose simple patterns on the behavior of matter in the
universe.
Furthermore, on the basis of our knowledge of physics, we cannot
rule out the appearance of highly complex patterns reflecting the willful
exercise of higher intelligence. Such patterns might exhibit complex regu
larities representing the execution of higher-order laws, or they might cor
respond to unique and seemingly capricious displays of free will. As we
have pointed out in Chapters 3 and 5, scientists actually know very little
about the laws of nature, and they have found it very difficult to apply the
known laws to complex phenomena. Our present knowledge is certainly
compatible with the idea that nature is directed by higher intelligence.
Here the objection might be raised that if all the phenomena of the uni
verse are arising from the will of a supremely intelligent being, then why
do so many of these phenomena appear chaotic and meaningless? Part
of the answer to this question is that meaningful patterns may appear ran
dom if they contain a high density of information. We have seen that such
Notes 165
complex patterns tend to obey certain statistical laws simply as a conse

quence of their large information content. Thus a complex, seemingly ran
dom pattern in nature may actually be meaningful, even though we do not
understand it. Another part of the answer is that meaningless patterns can
easily be generated by the transformation of meaningful patterns. This is
illustrated by the fact that many meaningful conversations, when heard
simultaneously in a crowded room, merge together into a meaningless din.
Such meaningless patterns will inherit the statistical properties of their
meaningful sources, and will appear as undecipherable "random noise."
These considerations enable us to understand how apparent meaning-
lessness and chaos can arise in nature, but they tell us nothing about the
definition of meaning itself. Thus far, no one has been able to give a satis
factory definition of meaning or purpose within the framework of the
mechanistic world view. However, such a definition is provided by the
Bhagavad-gītā. As we shall see, the ultimate meaning of the universe can
be understood by first understanding the nature of the individual conscious
beings, and the nature of their relationships with both material nature and
the Supreme Person. We will discuss these matters step-by-step in the
chapters that follow.
Notes
1. Bell, Men of Mathematics, p. 172.
2. We should note that for large n, the n u m b e r of possible subsequences,

2", becomes much larger than the length of o u r sequence, and so most
of these subsequences will not appear at all. This means that when we
deal with a finite a m o u n t of data, as we must in practice, we have no
basis for saying definitely whether or not a sequence of events obeys a
statistical law.
M a n y different m e t h o d s of formulating statistical laws have been
studied, and here we have briefly introduced the most widely used of
these, which is known as the m e t h o d of relative frequency. For a de
tailed discussion of the most important m e t h o d s of defining proba
bilities and statistical laws, see Fine, Theories of Probability. After an
exhaustive analysis, Fine ascertains that n o n e of these formulations is
satisfactory, and he concludes that, "Conceivably, probability is not
possible." (p. 248).
3. G n e d e n k o , Theory of Probability, p. 55.

4. Weinberg, " T h e Forces of N a t u r e , " p. 175.
5. Shannon, "A Mathematical Theory of C o m m u n i c a t i o n , " p. 379.

6. This sequence was encoded using the m e t h o d of Huffman, "A M e t h o d
for the Construction of Minimum Redundancy C o d e s , " p. 1098. As it
stands the sequence is highly r a n d o m , but since it still contains the re-
dundancy caused by the repetition of words such as "evolution," it is
not fully so. Thus further compression and consequent randomization
are possible.
7. Dobzhansky, "From Potentiality to Realization in Evolution," p. 20.

8. We should note that in his article Dobzhansky does not clearly define
his conception of the ultimate principles underlying the p h e n o m e n a of
the universe. He says that evolution is not acausal, that it is not due to
pure chance, and that it is due to many interacting causal chains. Yet he
also says that evolution is not rigidly predestined. He says that the
course of evolution was not p r o g r a m m e d or encoded into the primor-
dial universe, but that primordial matter had the potential for giving
rise to all forms of life, including innumerable unrealized forms. He
speaks of evolution in terms of probabilities and stresses that the
probability of the development of life as we know it is zero.
It appears that Dobzhansky is thinking in terms of causal inter-
actions that include, at some point, some mysterious element of abso-
lute chance. We can only conclude that his thinking is muddled. We
suggest that the reason for this is that although he needs the concept of
absolute chance in order to formulate his evolutionary world view, at
the same time he recognizes the illogical nature of this concept and
would like to avoid it. T h u s , he is caught in a dilemma.
9. Darwin, The Life and Letters of Charles Darwin, p. 20.

10. A s s u m e that the diameter of the spheres is ¼ inch. If there is an aver-
age m o v e m e n t of about 2 inches between bounces, then a slight varia-
tion in direction will be magnified by an average of at least 16 times per
bounce. Consequently, an error in the nth decimal place in the direc-
tion of motion will begin to affect the first decimal place after about
.83n bounces.
11. We should note that the unpredictable behavior of the bouncing ball
in our example is a c o m m o n feature of many nonlinear dynamical
models. See, for example, Ruelle, "Strange A t t r a c t o r s , " p p . 126-137.
12. T h e relation between computer operations and self-referential con-
sciousness can be further elucidated by the following thought experi-
Notes 167
m e n t . Suppose that a given c o m p u t e r , by virtue of its programming,

is actually able to experience the consciousness of Joe Smith con
templating a sunny s u m m e r afternoon. We might imagine that the
c o m p u t e r is able to do this by rapidly going through various loops rep
resenting J o e ' s awareness of his own awareness. T h e question is: W h a t
h a p p e n s if the c o m p u t e r is slowed d o w n , or even "single s t e p p e d " by
the o p e r a t o r at a very slow rate? T h e p a t t e r n of c o m p u t e r operations
remains t h e same, but is t h e same consciousness still t h e r e , stretched
out over time? Is t h e same consciousness still " t h e r e " if the c o m p u t e r is
m a d e to execute o n e step per year? T h e absurdity of this is an indica
tion that consciousness cannot be represented in terms of c o m p u t e r
operations.
13. A . C . B h a k t i v e d a n t a Swami P r a b h u p ā d a , Śrī īśopanisad, p. 1.
14. Bhakti Siddhanta Saraswati Thakur, Shri Brahma Samhita, p p . 99-100.
15. Schrodinger, What Is Life? and Mind and Matter, p. 93.
16. Schrodinger, p. 93.
17. This point is discussed in detail, with many references to t h e history

of Indian philosophy, in Kavirāja, Śrī Caitanya-caritāmrta, Ādi-līlā,
Vol. 2, chap. 7.
Chapter 7
On Inspiration
In this chapter we will examine how human beings acquire knowledge in
science, mathematics, and art. Our focus shall primarily be on the forma
tion of ideas and hypotheses in science and mathematics, since the formal
nature of these subjects tends to put the phenomena we are concerned
with into particularly clear perspective. We will show that the phenomenon
known as inspiration plays an essential part in acquiring knowledge in mod
ern science and mathematics and the creative arts (such as music). We will
argue that the phenomenon of inspiration cannot readily be explained
by mechanistic models of nature consistent with present-day theories of
physics and chemistry.
To give a positive alternative to these models, we will continue the proj
ect of outlining the nonmechanistic theoretical system of the Bhagavad-
gītā. Thus far we have introduced the concept of the conscious self or
jīvātmā as an entity distinct from the material body (Chapters 1 and 2),
and we have also introduced the concept of the all-pervading supercon-
scious being or paramātmā. The Bhagavad-gītā shows how the idea of the
paramātmā can be used to construct a model of the interaction between the
conscious self and the material body. As we shall see, this model accounts
for the phenomenon of inspiration in a direct and striking way.
Modern scientists acquire knowledge, at least in principle, by what is
called the hypothetico-deductive method. Using this method, they formu
late hypotheses and then test them by experimental observation. Inves
tigators consider hypotheses valid only insofar as they are consistent with
the data obtained by observation, and they must in principle reject any
hypothesis that disagrees with observation. Much analysis has been di
rected toward the deductive side of the hypothetico-deductive method, but
the equally important process of hypothesis formation has been largely
neglected. So we ask, "Where do hypotheses come from?"
It is clear that scientists cannot use any direct, step-by-step process to
derive hypotheses from raw observational data. To deal with such data at
all, they must already have some working hypothesis, for otherwise the
data amount to nothing more than a bewildering array of symbols (or sights
and sounds), which is no more meaningful than a table of random numbers.
In this connection Albert Einstein once said, "It may be heuristically useful
to keep in mind what one has observed. But on principle it is quite wrong to
170 On Inspiration
try grounding a theory on observable magnitudes alone. In reality the op-

1
posite happens. It is the theory which determines what we can observe."
Pure mathematics contains an equivalent of the hypothetico-deductive
method. In mathematics, instead of hypotheses there are proposed systems
of mathematical reasoning intended to answer specific questions. A n d in-
stead of the experimental testing of a hypothesis there is the step-by-step
process of verifying that a particular proof, or line of mathematical reason-
ing, is correct. This verification process is straightforward and could in
principle be carried out by a computer. However, there is no systematic,
step-by-step method of generating mathematical proofs and systems of
ideas, such as group theory or the theory of Lebesque integration.
If hypotheses in science and systems of reasoning in mathematics are not
generated by any systematic procedure, then what is their source? We find
that they almost universally arise within the mind of the investigator by
sudden inspiration. The classic example is Archimedes' discovery of the
principle of specific gravity. The Greek mathematician was faced with the
task of determining whether a king's crown was solid gold without drilling
any holes in it. After a long period of fruitless endeavor, he received the
answer to the problem by sudden inspiration while taking a bath.
Such inspirations generally occur suddenly and unexpectedly to per-
sons who had previously made some unsuccessful conscious effort to solve
the problem in question. They usually occur when one is not consciously
thinking about the problem, and they often indicate an entirely new way
of looking at i t — a way the investigator had never even considered during
his conscious efforts to find a solution. Generally, an inspiration appears
as a sudden awareness of the problem's solution, accompanied by the con-
viction that the solution is correct and final. One perceives the solution in
its entirety, though it may be quite long and complicated when written out
in full.
Inspiration plays a striking and essential role in the solution of difficult
problems in science and mathematics. Generally, investigators can success-
fully tackle only routine problems by conscious endeavor alone. Significant
advances in science almost always involve sudden inspiration, as the lives
of great scientists and mathematicians amply attest. A typical example is
the experience of the nineteenth century mathematician Karl Gauss. After
trying unsuccessfully for years to prove a certain theorem about whole
numbers, Gauss suddenly became aware of the solution. He described his
experience as follows: "Finally, two days ago I succeeded . . . Like a sud-
den flash of lightning, the riddle happened to be solved. I myself cannot say
what was the conducting thread which connected what I previously knew
2
with what made my success possible."
On Inspiration 171
We can easily cite many similar examples of sudden inspiration. Here is

another one, given by Henri Poincare, a famous French mathematician
of the late nineteenth century. After working for some time on certain
problems in the theory of functions, Poincare had occasion to go on a geo-
logical field trip, during which he set aside his mathematical work. While
on the trip he received a sudden inspiration involving his researches, which
he described as follows: " A t the moment when I put my foot on the step the
idea came to me, without anything in my former thoughts seeming to have
paved the way for it, that the transformations I had used . . . were iden-
3
tical with those of non-Euclidean geometry." Later on, after some fruit-
less work on an apparently unrelated question, he suddenly realized, "with
just the same characteristics of brevity, suddenness, and immediate cer-
4
tainty," that this work could be combined with his previous inspiration to
provide a significant advance in his research on the theory of functions.
Then a third sudden inspiration provided him with the final argument he
needed to complete that work.
Although inspirations generally occur after a considerable period of
intense but unsuccessful effort to consciously solve a problem, this is not
always the case. Here is an example from another field of endeavor. Wolf-
gang Mozart once described how he created his musical works: "When
I feel well and in good humor, or when I am taking a drive or walking, . . .
thoughts crowd into my mind as easily as you could wish. Whence and how
do they come? I do not know and I have nothing to do with it. . . Once I
have a theme, another melody comes, linking itself with the first one, in
accordance with the needs of the composition as a whole . . . Then my soul
is on fire with inspiration, if however nothing occurs to distract my atten-
tion. The work grows; I keep expanding it, conceiving it more and more
clearly until I have the entire composition finished in my head, though it
may be long . . . It does not come to me successively, with its various parts
worked out in detail, as they will be later on, but it is in its entirety that my
5
imagination lets me hear i t . " (Italics added.)
From these instances we discover two significant features of the phe-
nomenon of inspiration: first, its source lies beyond the subject's conscious
perception; and second, it provides the subject with information unob-
tainable by any conscious effort. These features led Poincare and his fol-
lower Hadamard to attribute inspiration to the action of an entity which
Poincare called the "subliminal self," and which he identified with the sub-
conscious or unconscious self of the psychoanalysts. Poincare made the fol-
lowing interesting observations about the subliminal self: "The subliminal
self is in no way inferior to the conscious self; it is not purely automatic; it is
capable of discernment; it has tact, delicacy; it knows how to choose, to
172 On Inspiration
divine. What do I say? It knows better how to divine than the conscious
self, since it succeeds where that has failed. In a word, is not the subliminal
6
self superior to the conscious self?" Having raised this question, Poincare
then backs away from it: "Is this affirmative answer forced upon us by the
facts I have just given? I confess that for my part, I should hate to accept
7
i t . " He then offers a mechanical explanation of how the subliminal self,
viewed as an automaton, could account for the observed phenomena of
inspiration.
7.1 The Mechanistic Explanation

Let us carefully examine the arguments for such a mechanical explana-
tion of inspiration. This issue is of particular importance at the present
time, because the prevailing materialistic philosophy of modern science
holds that the mind is nothing more than a machine, and that all mental
phenomena, including consciousness, are nothing more than the products
of mechanical interactions. The mental machine is specifically taken to be
the brain, and its basic functional elements are believed to be the nerve
cells and possibly some systems of interacting macromolecules within these
cells. Many modern scientists believe that all brain activity results simply
from the interaction of these elements according to the known laws of
physics.
No scientist has yet formulated an adequate explanation of the differ-
ence between a conscious and an unconscious machine, or even indicated
how a machine could be conscious at all. In fact, investigators attempting to
describe the self in mechanistic terms concentrate exclusively on the dupli-
cation of external behavior by mechanical means; they totally disregard the
individual person's subjective experience of conscious self-awareness. (See
Chapter 2.) This approach to the self is characteristic of modern behavioral
psychology. It was formally set forth by the British mathematician Alan
Turing, who argued that since whatever a human being can do a computer
can imitate, a human being is merely a machine.
For the moment we will follow this behavioristic approach and simply
consider the question of how the phenomenon of inspiration could be du-
plicated by a machine. Poincare proposed that the subliminal self must put
together many combinations of mathematical symbols by chance until at
last it finds a combination satisfying the desire of the conscious mind for a
certain kind of mathematical result. He proposed that the conscious mind
would remain unaware of the many useless and illogical combinations run-
ning through the subconscious, but that it would immediately become
aware of a satisfactory combination as soon as it was formed. He therefore
The Mechanistic Explanation 173
proposed that the subliminal self must be able to form enormous numbers
of combinations in a short time, and that these could be evaluated sub-
consciously as they were formed, in accordance with the criteria for a
satisfactory solution determined by the conscious mind.
As a first step in evaluating this model, let us estimate the number of
combinations of symbols that could be generated within the brain within a
reasonable period of time. A very generous upper limit on this number is
46
given by the figure 3.2 x 10 . We obtain this figure by assuming that in
each cubic Angstrom unit of the brain, a separate combination is formed
and evaluated once during each billionth of a second over a period of one
hundred years. Although this figure is an enormous overestimate of what
the brain could possibly do within the bounds of our present understanding
of the laws of nature, it is still infinitesimal compared to the total number of
possible combinations of symbols one would have to form to have any
chance of randomly hitting a proof for a particular mathematical theorem
of moderate difficulty.
If we attempt to elaborate a line of mathematical reasoning, we find that
at each step there are many possible combinations of symbols we can write
down, and thus we can think of a particular mathematical argument as a
path through a tree possessing many successive levels of subdividing
branches. This is illustrated in the figure below. The number of branches in
such a tree grows exponentially with the number of successive choices, and
the number of choices is roughly proportional to the length of the argu-
ment. Thus as the length of the argument increases, the number of
46 10(l
branches will very quickly pass such limits as 1 0 and 10 . For example,
suppose we are writing sentences in some symbolic language, and the rules
of grammar for that language allow us an average of two choices for each
Figure 1. T h e relationship between different possible lines of mathematical rea-

soning can be represented by a tree. Each node represents a choice a m o n g various
possibilities that restricts t h e further d e v e l o p m e n t of the a r g u m e n t .
i
174 On Inspiration
100
successive symbol. Then there will be approximately 1 0 grammatical
sentences of 333 symbols in length.
Even a very brief mathematical argument will often expand to great
length when written out in full, and many mathematical proofs require
pages and pages of highly condensed exposition, in which many essential
steps are left for the reader to fill in. Thus there is only an extremely remote
chance that an appropriate argument would appear as a random combina-
tion in Poincare's mechanical model of the process of inspiration. Clearly,
the phenomenon of inspiration requires a process of choice capable of
going more or less directly to the solution, without even considering the
vast majority of possible combinations of arguments.
7.2 Some Striking Examples
The requirements that this process of choice must meet are strikingly il-
lustrated by some further examples of mathematical inspiration. It is very
often found that the solution to a difficult mathematical problem depends
on the discovery of basic principles and underlying systems of mathemati-
cal relationships. Only when these principles and systems are understood
does the problem take on a tractable form. Therefore difficult problems
have often remained unsolved for many years, until mathematicians <
gradually developed various sophisticated ideas and methods of argu-
ment that made their solution possible. However, it is interesting to note
that on some occasions sudden inspiration has completely circumvented
this gradual process of development. There are several instances in which
famous mathematicians have, without proof, stated mathematical results
that later investigators proved only after elaborate systems of underlying
relationships had gradually come to light. Here are two examples:
The first example concerns the zeta-function studied by the German
mathematician Bernhard Riemann. At the time of his death, Riemann left
a note describing several properties of this function that pertain to the
theory of prime numbers. He did not give any proof for existence of these
properties, and many years elapsed before mathematicians were able to
find existence proofs for all but one of them. The remaining question is still
unsettled, though an immense amount of labor has been devoted to it over
the last seventy-five years. Of the properties of the zeta-function that have
been verified, the mathematician Jacques Hadamard said, " A l l these com-
plements could be brought to Riemann's publication only by the help of
facts which were completely unknown in his time; and, for one of the prop-
erties enunciated by him, it is hardly conceivable how he can have found it
Some Striking Examples 175
without using some of these general principles, no mention of which is

8
made in his paper."
The work of the French mathematician Evariste Galois provides us with
a case similar to Riemann's. Galois is famous for a paper, written hurriedly
in sketchy form just before his death, that completely revolutionized the
subject of algebra. However, the example we are considering here con-
cerns a theorem Galois stated, without proof, in a letter to a friend. Ac-
cording to Hadamard this theorem could not even be understood in terms
of the mathematical knowledge of that time; it became comprehensible
only years later, after the discovery of certain basic principles. Hadamard
remarks "(1) that Galois must have conceived these principles in some way;
(2) that they must have been unconscious in his mind, since he makes no al-
9
lusion to them, though they themselves represent a significant discovery."
It would appear, then, that the process of choice underlying mathemati-
cal inspiration can make use of basic principles that are very elaborate and
sophisticated and that are completely unknown to the conscious mind of
the person involved. Some of the developments leading to the proof of
Riemann's theorems are highly complex, requiring many pages (and even
volumes) of highly abbreviated mathematical exposition. It is certainly
hard to see how a mechanical process of trial and error, such as that de-
scribed by Poincare, could exploit such principles. Yet, if other, simpler
solutions exist that avoid the use of such elaborate developments, they
have remained unknown up to the present time, despite the extensive
research devoted to these topics.
The process of choice underlying mathematical inspiration must also
make use of selection criteria that are exceedingly subtle and hard to de-
fine. Mathematical work of high quality cannot be evaluated simply by the
application of cut-and-dried rules of logic. Rather, its evaluation involves
emotional sensibility and the appreciation of beauty, harmony, and other
delicate aesthetic qualities. Of these criteria Poincare said, " I t is almost im-
10
possible to state them precisely; they are felt rather than f o r m u l a t e d . "
This is also true of the criteria by which we judge artistic creations, such as
musical compositions. These criteria are very real but at the same time very
difficult to define precisely. Yet evidently they were fully incorporated in
that mysterious process which provided Mozart with sophisticated musical
compositions without any particular effort on his part and, indeed, without
any knowledge on his part of how it was all happening.
If the process underlying inspiration is not one of extensive trial and
error, as Poincare suggested, but rather one that depends mainly on direct
choice, then we can explain it in terms of current mechanistic ideas only by
176 On Inspiration
positing the existence of a very powerful algorithm (a system of computa-

tional rules) built into the neural circuitry of the brain. However, it is not at
all clear that we can satisfactorily explain inspiration by reference to such
an algorithm. Here we will briefly discuss two major problems entailed by
the brain-algorithm hypothesis.
(1) Origins. If mathematical, scientific, and artistic inspirations result
from the workings of a neural algorithm, then how does the pattern of
nerve connections embodying this algorithm arise? We know that the
algorithm cannot be a simple one when we consider the complexity of the
automatic theorem-proving algorithms that have been produced thus
far by workers in the field of artificial intelligence.'' These algorithms can-
not even approach the performance of advanced human minds, and yet
they are extremely elaborate. But if our hypothetical brain-algorithm is
extremely complex, how did it come into being? It can hardly be accounted
for by extensive random genetic mutation or recombination in a single
generation, for then the problem of random choice among vast numbers
of possible combinations would again arise. One would therefore have
to suppose that only a few relatively probable genetic transformations
separated the genotype of Mozart from those of his parents, who, though
talented, did not possess comparable musical ability.
However, it is not the usual experience of those who work with algo-
rithms that a few substitutions or recombinations of symbols can drastically
improve an algorithm's performance or give it completely new capacities
that would impress us as remarkable. Normally, if this were to happen with
a particular algorithm, we would tend to suppose that it was a defective ver-
sion of another algorithm originally designed to exhibit those capacities.
This would imply that the algorithm for Mozart's unique musical abilities
existed in a hidden form in the genes of his ancestors.
This brings us to the general problem of explaining the origin of human
traits. According to the theory most widely accepted today, these traits
were selected on the basis of the relative reproductive advantage they con-
ferred on their possessors or their possessors' relatives. Most of the selec-
tion for our hypothetical hidden algorithms must have occurred in very
early times, because of both the complexity of these algorithms and the fact
that they often must be carried in a hidden form. It is now thought that
human society, during most of its existence, was on the level of hunters and
gatherers, at best. It is quite hard to see how, in such societies, persons like
Mozart or Gauss would ever have had the opportunity to fully exhibit their
unusual abilities. But if they didn't, then the winnowing process posited by
the theory of evolution could not effectively select these abilities.
The Interaction Between Consciousness and Matter 177
We are thus faced with a dilemma: It appears that it is as difficult to ac

count for the origin of our hypothetical inspiration-generating algorithms
as it is to account for the inspirations themselves.
(2) Subjective experience. If the phenomenon of inspiration is caused
by the working of a neural algorithm, then why is it that an inspiration
tends to occur as an abrupt realization of a complete solution, without the
subject's conscious awareness of intermediate steps? The examples of
Riemann and Galois show that some persons have obtained results in an
apparently direct way, while others were able to verify these results only
through a laborious process involving many intermediate stages. Nor
mally, we solve relatively easy problems by a conscious, step-by-step pro
cess. Why, then, should inspired scientists, mathematicians, and artists
remain unaware of important intermediate steps in the process of solving
difficult problems or producing intricate works of art, and then become
aware of the final solution or creation only during a brief experience of
realization?
Thus we can see that the phenomenon of inspiration cannot readily be
explained by means of mechanistic models of life consistent with present-
day theories of physics and chemistry. In the remainder of this chapter we
will explore the nonmechanistic approach of the Bhagavad-gītā.
7.3 The Interaction Between Consciousness and Matter

In the first part of this book we discussed consciousness and, drawing
from the Bhagavad-gītā, introduced the conception of the conscious self as
a nonphysical entity. In the second part we discussed biological form and
the problem of finding a unified description of nature. We concluded that
it is not possible to construct a unified quantitative theory that can give
a satisfactory account of the origin of complex forms, such as the bodily
structures of living organisms. But turning once again to the Bhagavad-
gītā, we showed how an intuitively reasonable, unified picture of nature
can be based on the nonquantifiable idea of universal consciousness.
Thus far we have not given any indication of how the conscious self, or
jīvātmā, influences the behavior of the material body. As we observed in
Chapter 2, philosophers have not been able to solve this problem, and
many have responded either by denying the existence of consciousness or
by trying to explain it as a byproduct of brain activity. Nonetheless, a sim
ple solution does exist. The Bhagavad-gītā indicates that the interplay be
tween the conscious self and the material body is indirect, and depends on
the interaction between the localized self and universal consciousness. This
explanation of the relationship between the jīvātmā and matter resolves
178 On Inspiration
many of the problems that have perplexed philosophers, and it also directly
accounts for the phenomena of inspiration. Most importantly, this expla
nation entails a direct method of obtaining verifiable knowledge about
both localized and universal consciousness.
The description of the interaction between the jīvātmā and the material
body given in the first part of the Bhagavad-gītā may seem perplexing:
"The bewildered spirit soul, under the influence of the three modes of ma
terial nature, thinks himself to be the doer of activities, which are in actual
ity carried out by nature." 1 2 Or again: "The embodied spirit, master of the
city of his body, does not create activities, nor does he influence people to
act, nor does he create the fruits of action. A l l this is enacted by the modes
of material nature." 1 3 Apparently, these statements support the viewpoint
that nature is working entirely in accordance with certain fixed laws, and
that the conscious self can at best be an epiphenomenon affected by the
actions of the material body but unable to affect them in turn.
The Bhagavad-gītā confirms that nature is indeed working according to
laws. But the key to the solution of the mind-body problem lies in under
standing the character of these laws. Physicists tend to conceive of the laws
of nature as a closed set of immutable rules that can be specified by a few
simple equations. According to the Bhagavad-gītā, however, the laws of
nature are like the laws of human society promulgated by a head of state.
Like the laws of the physicists, these laws can also be represented in sym
bolic terms, but unlike them they tend to be highly complex. A n d since the
laws of nature are actually under personal control, they are always subject
to interference and modification.
In the Bhagavad-gītā the Supreme Person, Krsna, describes the laws of
nature in the following terms: "This material nature is working under My
direction, O son of Kuntī, and it is producing all moving and nonmoving
beings. By its rule this manifestation is created and annihilated again and
again." 1 4 Thus although the physicists are right that the material energy is
acting according to rules, they fail to see that these rules have their ultimate
source in a personal director.
We can visualize these rules in terms of a hierarchy. On the lowest level
are the relatively simple laws governing the gross behavior of matter.
These are studied to some extent by physicists and chemists. On the next
level are higher-order laws governing the complex behavior of the living
beings. These laws, sometimes referred to as the laws of karma (action),
are discussed in the Bhagavad-gītā in some detail. Finally, on the highest
level are the direct interventions of the Supreme Person in the course of
natural events. We have discussed in Parts I and II how one can extend
the world view of modern science to encompass such an open»ended hier-
The Interaction Between Consciousness and Matter 179
archy of natural laws. In such a hierarchy the laws on each level are not, of
course, absolute. They are merely approximations subject to refinement
and modification in accordance with higher laws and, ultimately, the
unconstrained will of the Supreme Person.
From the mechanistic viewpoint, such an "unconstrained w i l l " is at best
nothing more than a name for the arbitrary and the inexplicable. Yet the
Bhagavad-gītā describes some additional nonquantifiable factors that can
give us greater insight into the will of the Supreme Person. One of these
factors is the personal interaction between the Supreme and the localized
conscious selves.
In the Bhagavad-gītā Krsna states, "The Supreme Lord is situated in
everyone's heart, O A r j u n a , and is directing the wanderings of all living
entities, who are situated on a machine made of the material energy." 1 5 As
we have already mentioned, the materially embodied jīvātmā is in an
essentially passive state, unable directly to influence the actions of the .
material body. Yet the jīvātmā is conscious of the bodily situation and filled
with desires relating to the outcome of various bodily activities. According
to the Bhagavad-gītā, the Supreme Person perceives the desires of the e m - '
bodied beings and responds to these desires by appropriately controlling
the bodily machinery.
The manifestation of the Supreme Person within the heart of every living
being is known as the paramātmā, or Supersoul. "Although the Supersoul
appears to be divided," says the Bhagavad-gītā, "He is never divided. He
is situated as one." 1 6 This is another illustration of the simultaneous one
ness and multiplicity of the Supreme Person, which we discussed in Chap
ter 6. Since the Supreme Person possesses unlimited consciousness, He is
able to attend simultaneously to innumerable material situations without
becoming confused.
The Supreme Person, as understood from the Bhagavad-gītā, is not
remote from the material world. Rather, He is all-pervading in space and
time and is also transcendental to space and time. This idea may seem para
doxical, but we should note that a similar problem arises when we try to
visualize the reality underlying the laws of nature as conceived in modern
physics. These laws are postulated as spatially and temporally invariant;
but what is it that pervades all space and time and determines that gravita
tion, for example, will operate according to a certain universal force
constant, G?
In our day-to-day experience we desire to perform various physical ac
tions, and we generally find that the body immediately acts in accordance
with our desires. Although we do not understand how our will gives rise to
action, this seems to happen automatically, and we normally take it for
180 On Inspiration
granted and think, "I am doing this." According to the Bhagavad-gītā,

what is actually happening is that the Supersoul is perceiving our desires
and translating them into action. He does this through the manipulation of
the laws of nature on a sophisticated level. As a consequence, our actions
seem to conform to the known physical laws, even though, if we could
analyze these actions thoroughly enough, we would find no fixed system of
laws ultimately able to account for them.
The philosophy of the Bhagavad-gītā thus provides us with a simple ex
planation of the phenomenon of inspiration—the very explanation, alas,
that Poincare, after analyzing his own mathematical inspirations, felt he
would hate to accept. In each of the cases we have discussed, a person ini
tially desired very strongly to carry out a difficult mental task, such as find
ing the proof of a mathematical theorem. Generally, the person spent some
time fruitlessly endeavoring to execute the task, and then he suddenly and
unexpectedly became aware of the solution to his problem. (The example
of Mozart shows, however, that an initial period of frustrating failure is not
always involved.) We have seen that in many instances such unexpected
realizations cannot be satisfactorily explained as byproducts of known
physical processes. But we can readily understand experiences of sudden il
lumination as the responses of the Supersoul to the desires of the embodied
living being.
We should note that different persons will generally have different ex
periences of inspiration. Even if two people desiring to solve the same
problem are of similar ability and education, one may find the solution
and the other may not. The Bhagavad-gītā explains this variance as a con
sequence of karma, or the accumulated byproducts of past actions. The
Supersoul does not direct the actions of the material body solely in accor
dance with the immediate desires of the living being. Rather, He deter
mines these actions by systematically evaluating both the living being's
current desires and his past activities. He bases these judgements on an
established set of higher-order laws—the laws of karma—which provide
an absolute standard for universal justice and morality. These laws are just
as real as the "natural laws" of physics and chemistry. But they are much
more complex than these laws, and they are directly involved with the
phenomenon of life.
The Bhagavad-gītā deals primarily with the means whereby the indi
vidual conscious being can become free from the bondage of past karma.
A living being acting under the laws of karma is more or less a helpless
spectator of an elaborate drama of actions and reactions stemming from
past and present desires. In this state the living being tends to identify fully
with his physical body and to be completely unaware of the role played by
Notes 181
the Supersoul in the affairs of his life.

Under the laws of karma, the relationship between the individual
jīvātmā and the Supreme Person is essentially impersonal and legalistic.
Yet it is possible for the jīvātmā to become directly aware of the Supreme
Person and reciprocate with H i m in a personal relationship. In such a
state of consciousness, the jīvātmā is freed from karmic reactions to past
activities and becomes fully aware of his own nature as a nonphysical being.
Inspiration plays an essential role in the attainment of this state of con
sciousness. In the Bhagavad-gītā Krsna declares, "To those who are con
stantly devoted and worship Me with love, I give the understanding by
which they can come to Me. Out of compassion for them, I dwelling in their
hearts, destroy with the shining lamp of knowledge the darkness born of
17
ignorance." The purport of this statement is that direct knowledge of
the Supreme Person is available to any person willing to approach the
Supreme with a positive, favorable attitude. Normally, Krsna supplies
the embodied jīvātmā only with information relating to his particular ma
terial desires, but if the jīvātmā approaches Krsna with love and without
underlying material motives, then Krsna will directly reveal Himself.
This is the conclusion of the Bhagavad-gītā's philosophy, and it con
stitutes the only real means of verifying the truth of this philosophy. By
analyzing empirical arguments, we can identify some of the shortcomings
of mechanistic theories and show how the philosophy of the Bhagavad-gītā
can supply important elements now missing from the prevailing scientific
world view. But if we remain within the framework of mechanistic thought,
we can neither prove that this philosophy is true nor practically apply it. We
can verify transcendental subject matter only if we are able to actually
function on the transcendental platform.
In the discussion thus far we have indicated that according to the
Bhagavad-gītā, such transcendental consciousness is theoretically attain
able. The individual self is always in contact with the Supersoul and is
capable, in principle, of relating with the Supersoul on the level of direct
personal exchange. In Chapter 9 we will discuss the epistemology.of tran
scendental knowledge and indicate briefly how this theoretical possibility
might be practically realized.
Notes
1. Brush, "Should the History of Science be R a t e d X ? " p. 1167.
2. H a d a m a r d , The Psychology of Invention in the Mathematical Field,

p . 15.
182 On Inspiration
3. Poincare, The Foundations of Science, pp. 387-388.
4. Ibid.
5. H a d a m a r d , p. 16.
6. Poincare, p. 390.
8. H a d a m a r d , p. 118.
9. H a d a m a r d , p. 120.
11. Weizenbaum, Computer Power and Human Reason, chap. 9.
12. A . C . Bhaktivedanta Swami Prabhupāda, Bhagavad-gītā As It Is,

text 3.27, p. 192.
13. Ibid., text 5.14, p. 286.
14. Ibid., text 9.10, p. 457.
15. Ibid., text 18.61, p. 830.
16. Ibid., text 13.17, p. 640.
17. Ibid., texts 10.10-11, p p . 506-508.

Chapter 8
The Doctrine
Of Evolution
" A l l reputable evolutionary biologists now agree that the evolution of
1
life is directed by the process of natural selection, and by nothing else."
With these words Sir Julian Huxley summed up the consensus of learned
opinion at the Darwin Centennial Celebration in 1959.
Among the eminent biologists and evolutionists attending the cele-
bration , great confidence prevailed that the origin of living species was now
almost fully understood. Evolutionists had clearly established that all
living organisms had gradually evolved through small variations in form
and function, slowly accumulating, generation by generation, over a vast
span of geological time. Geneticists had shown that all biological varia-
tions arose from random genetic accidents called mutations. Evolutionary
theorists, building on this finding, had clearly identified Darwinian natural
selection as the sole guiding force that sorted out these variations and
thereby molded the diverse forms of living beings. Although many minute
details certainly remained to be worked out, scientists believed they had
arrived at an essentially complete understanding of life and its historical
development.
With this striking unanimity of established scientific opinion reached lit-
tle more than two decades ago, perhaps we are surprised to hear that the
theory of evolution has recently become the focus of a great controversy
among evolutionists themselves. The last few years have seen the estab-
lished theory of mutation and natural selection increasingly challenged by
critical studies and dissenting interpretations of the evidence. The theory
has clearly shown itself unsound, although scientists have thus far been
unable to devise an acceptable new theory to replace it.
Recently this controversy became a near battle, when some 150 promi-
nent evolutionists gathered at Chicago's Field Museum of Natural History
to thrash out various conflicting hypotheses about the nature of evolution.
After four days of heated discussions (closed to all but a few outside ob-
servers), the evolutionists remained convinced that evolution is a fact.
Unfortunately, however, they could not reach a clear understanding of
just what this fact is. According to a report from The New York Times, the
183
184 The Doctrine of Evolution
assembled scientists were unable either to specify the mechanism of evolu-

tion or to agree on "how anyone could establish with some certainty that it
2
happened one way and not another."
Why this shift from unanimity and certainty to controversy and indeci-
sion? We shall try to answer this question by examining some basic features
of the modern theory of evolution. We shall try to identify the reasons why
many scientists have sought an evolutionary explanation of life, and we
shall also point out some of the problems that have impeded their efforts.
Finally, we shall argue that the theory of evolution has been motivated
more by philosophical misunderstanding than by the strength of empirical
evidence. One of the most persistent themes of evolutionary argumenta-
tion has been that evolution is the only viable alternative to certain stereo-
typed conceptions of divine creation. The theory of evolution can thus be
seen as an effort to enlist natural evidence in support of an alternative to an
unwanted and unsatisfactory spiritual world view. We propose that this
effort is futile, and that the underlying philosophical problems can be
solved only if we take advantage of a direct, scientific approach to spiritual
knowledge.
8.1 Evolution: an Invisible Process

When Charles Darwin originally set forth his evolutionary theory, he
maintained that the forms of living organisms change slowly and continu-
ously from generation to generation, and that over many millions of years
these changes bring about new species and higher categories of organisms.
One immediate implication of this theory was that the fossil record of
ancient plant and animal life should display a continuum of fossilized
life forms, ranging from the most primitive to the most advanced. Given
that organisms tend to leave occasional fossilized remains, scientists natu-
rally expected to find a petrified motion picture of evolutionary history
entombed in the earth's sedimentary rocks.
But in Darwin's time it was well known that the fossil record did not actu-
ally reveal such a picture. On the contrary, paleontologists had observed
that distinct plant and animal species tended to appear abruptly in the fos-
sil strata without recognizable antecedents. Each species remained es-
sentially unchanged throughout the strata bearing its fossilized remains.
Fossils yielded practically no evidence of gradual change from one species
to another.
Darwin admitted that the fossil evidence, far from supporting his theory,
seemed directly to contradict it. He responded by proposing that the fos-
sil record was drastically incomplete. The innumerable intermediate life
Evolution: an Invisible Process 185
forms required by his theory must have existed, but they had left no recog-
nizable traces in the fossil deposits known in his time. Darwin suggested
that further research would undoubtably uncover many of these missing
forms, and their discovery would vindicate his theory.
For many years orthodox evolutionary opinion has adhered to Darwin's
basic views. But dissenting voices have increasingly been heard. At the re-
cent meeting of evolutionists in Chicago, Niles Eldridge, a paleontologist
from the American Museum of Natural History in New York, declared,
3
"The pattern we were told to find for the last 120 years does not exist."
Despite intense effort, several generations of paleontologists have found
few examples in which one fossil species seems to transform gradually into
another, and some researchers say none at all have been found.
As a result, Eldridge, Steven J. Gould, and several other prominent
paleontologists now propose that species have not actually arisen by a slow
process of transformation. As an alternative, they have devised what they
4
call the theory of "punctuated equilibrium." According to this theory,
evolutionary changes take place in short bursts separated by long periods
during which the forms of living organisms remain static. A typical species
will arise from an earlier species in a "geological microsecond"—a period
of a few thousand years that appears like an instant from the multimillion-
year perspective of geological time. Also, a species will not arise through a
gradual modification of its parent population. Rather, it will arise when a
Figure 1. The Darwinian theory of evolution states that species have developed
gradually, as illustrated by the branching tree pattern (a). But the fossil record does
not substantiate this pattern of development. As a result, some paleontologists
have introduced the "punctuated equilibrium" model, shown in (b), in which the
transitions between species are officially invisible.
tiny group that has been isolated from the main population, perhaps by a
geographical barrier, is rapidly transformed.
One consequence of the theory of punctuated equilibrium is that it
makes the evolution of species officially invisible. On one hand, we cannot
expect the fossil record to show how a new species evolved, for the evolu-
tion takes place in a tiny population during a geological "microsecond." On
the other hand, we cannot expect to see a new species evolve within the re-
corded span of human history, for a geological microsecond of 10,000 to
50,000 years is still immensely long when measured in human lifetimes.
Of course, we may possibly observe small-scale changes in organisms,
like those produced through controlled breeding, or like the famous
change in color exhibited by the peppered moths of industrial England. Yet
such changes are known to be reversible, and at most they result in only
minor variations within a species. For example, settlers introduced domes-
ticated rabbits into Australia in 1788, and some escaped and flourished in
the wild. Despite the effects of breeding by humans, these domesticated
rabbits were still classed as rabbits, and today their descendants have
5
reverted to their ancestral form: they look exactly like wild rabbits.
Explaining superficial variations of this kind is not the* real problem con-
fronting evolutionists. The real problem is explaining how higher forms of
plants and animals have arisen from lower forms, and how these in turn
have arisen from inanimate matter. No large-scale transformations of this
kind have ever been observed within the brief span of human history. The
orthodox Darwinian theory maintained that such transformations should
be directly visible in the fossil record. But the theory of punctuated equilib-
rium says we should not expect even the fossil record to show these trans-
formations. In fact, the actual process that brings about new species of life
has always been invisible. Now, in the new theory propounded by Eldridge
and Gould, this process is held to be invisible even in principle.
8.2 The Fossil Record and the Origin of Higher Plants
Even though the transitional stages between species are officially invisi-
ble according to the punctuated equilibrium model, one might still wonder
whether or not the fossil record shows a large-scale progression from primi-
tive to advanced life forms that appears reasonably continuous. When we
examine a half-tone picture under a magnifying glass, we see an array of
disconnected dots; but when we view the picture at a distance, these dots
seem to merge together to form a continuous image. Similarly, we might
expect the disjointed succession of life forms in the fossil record to merge
The Fossil Record and the Origin of Higher Plants 187
into an unbroken evolutionary continuum when viewed from a sufficiently

broad perspective.
If we examine the published descriptions of the fossil record, however,
we find that this is not the case. The fossil record is characterized by abrupt
appearances and disappearances on every level of morphological classifica-
tion from species to phyla. It is not possible for us to see an evolutionary
continuum by standing back, so to speak, and taking into account only the
most significant features of organisms. Even such important general pat-
terns as the basic vertebrate body plan are seen to appear suddenly in the .
fossil record without recognizable antecedents. It would certainly be mis-
leading to claim that the fossil record demands an evolutionary interpre-
tation. Indeed, some features of this body of data can be made to appear
consistent with such an interpretation only by the introduction of artificial,
ad hoc hypotheses.
There are many examples of such features, but here we shall restrict our-
selves to considering only one—the paleontological history of the flower-
ing plants. (This history is represented graphically in Figure 2.) We will
briefly discuss the implications of this history after first introducing some
necessary background information about the geological time scale.
Geologists and paleontologists divide the history of the earth into a
series of time intervals that they correlate with successive strata of sedi-
mentary rock in the earth's surface. Table 1 summarizes the standard
6
geological time scale. Geological time is divided into successively finer
intervals called eras, periods, and epochs. Dates have been assigned to
these intervals by a complex process involving radiometric dating and vari-
ous methods of analyzing and correlating sedimentary strata. In Table 1 the
epochs making up each period are designated by numbers. These epochs
vary considerably in length, but their average length is about six million
years.
In Figure 2 the numbers along the horizontal axis represent the first 40
epochs in the geological time scale, and the triangles mark the boundaries
of the first five geological periods. For each epoch, we can read from the
graph the number of families of flowering plants that paleontologists have
7
found in the strata corresponding to that epoch.
The most striking feature of this graph is that in epochs 20 and 19, the
Cenomanian and Albian epochs of the Cretacious period, the number of
families rises abruptly from 0 to about 50. If we interpret the record liter-
ally, this means that at the end of the Cenomanian epoch (about 100 million
years ago), there existed hundreds of species of angiosperms, or flowering
plants, including many familiar modern trees and herbs. From this time
Period
boundaries
Epoch in millions
Era Period numbers of years
Cenozoic Quaterniary 1-2 0-2

Tertiary 3-13 2-65
Mesozoic Cretacious 14-25 65-136
Jurassic 26-35 136-190
Triassic 36-42 190-225
Paleozoic Permian 43-47 225-280
Carboniferous 48-53 280-345
Devonian 54-60 345-395
Silurian 61-63 395-435
Ordovician 64-69 435-500
Cambrian 70-72 500-570
Pre-Cambrian Varangian 73 . above 570
Pre-Varangian 74 —
Table 1. The geological time scale.
onward, the angiosperms have been the dominant land plants throughout
the world.
In contrast, in the strata antedating the Albian epoch paleontologists
have found very little unambiguous evidence for the existence of flowering
plants. There are some fossils, which suggest that angiosperms may have
8
existed in the Jurassic and Triassic periods. However, the compilers of
the standard reference used to draw Figure 2 apparently did not feel that
this evidence justified the inclusion of Jurassic or Triassic angiosperms in
their list of fossil families. In fact, according to this reference, two families
of angiosperms date from epochs 23 and 2 1 , and all other families appear in
epoch 20 or later.
Here we seem to have "punctuated equilibrium" with a vengeance. In a
span of some 12 million years, the flowering plants rise from a position of
complete insignificance to one of world dominance. Could this be due to an
exceedingly intense burst of evolutionary diversification? In fact, the gen-
eral opinion of paleontologists is that the angiosperms could not have
evolved so quickly. The prevailing view is that the angiosperms must have
gradually evolved in some part of the world where, for many millions of
years, they left no significant fossil remains. Then in the late Cretacious
Figure 2. T h e n u m b e r of families of flowering plants as a function of time. T h e num-

bers on the horizontal axis represent geological epochs. T h e graph depicts the
n u m b e r of families of flowering plants that paleontologists have discovered in the
strata corresponding to these epochs. In epochs 19 and 20 (the C e n o m a n i a n and
Albian divisions of the Cretacious period) the n u m b e r of families shoots up
abruptly from zero to about 50.
period they suddenly migrated en masse into the regions where their
remains are now found in abundance.
According to the paleontologist Daniel Axelrod, "The belief that early
angiosperm evolution took place in upland regions, in areas sufficiently re-
mote from lowland basins of deposition to have precluded their occurrence
9
in the record, is now generally conceded." Axelrod argues that upland re-
gions tend to be subjected to intense erosion, which quickly destroys any
fossils that might accumulate there. This implies that with the exception of
very recent highland deposits that have not yet been eroded, the highlands
of the world have not been represented in the existing fossil record.
Axelrod suggests that the entire evolutionary history of the flowering
plants took place in these unrecorded upland areas. He observes, "The an-
cestral group that gave rise to angiosperms has not yet been identified in
the fossil record, and no living angiosperm points to such an ancestral al-
10
liance." He also notes that the fossil record gives no indication of the
evolutionary relationship between different types of flowering plants: " I n
addition the record has shed almost no light on relations between taxa at
ordinal and family l e v e l . " " A l l evidence of these evolutionary relation-
ships was presumably obliterated by the erosion of the highlands in which
the early angiosperms exclusively lived.
Now, flowering plants may well have flourished in the highland regions
of the world prior to the Albian epoch of the Cretacious. However, it is
completely unscientific for paleontologists to try to save the theory of
evolution by maintaining that they evolved there. If this procedure is al-
lowed, then the theory of evolution becomes unfalsifiable. When evidence
for the evolution of a particular life form is lacking, it is easy to propose that
the missing evolutionary steps took place under circumstances that pre-
cluded the formation of a permanent record of the process. Such vacuous
proposals can explain anything, but for this very reason they have no place
in a scientific account of the world. Unfortunately, as we have already seen
in our discussion of the theory of punctuated equilibrium, it has become a
standard practice for paleontologists to claim that critical evolutionary
steps have occurred without leaving a record.
Now that we have identified one unscientific practice of the evolution-
ists, it may not be out of place here to point out a complementary practice
that is also unscientific. This is the procedure of setting aside and ignoring
evidence that does not fit into a particular evolutionary scheme. For exam-
ple, many paleontologists believe that the angiosperms must have had a
long evolutionary history prior to the Cretacious period, and thus they
would be willing to accept earlier fossil evidence of flowering plants. But
what would they say about evidence for the existence of flowering plants,
say two billion years ago?.
12
According to the standard scenario presented in textbooks and popu-
13
lar accounts , all higher vascular plants, including the angiosperms,
evolved from the psilophytes, an extremely primitive type of plant that
flourished in the upper Silurian and lower Devonian periods. The preced-
ing periods of the Paleozoic era are represented in the fossil record only by
marine deposits, and evolutionists generally suppose that life had not yet
ernerged from the seas during these periods. Prior to the Cambrian period,
the fossil record is very scanty, and evolutionists maintain that throughout
most of this time, life existed only in the form of single-celled organisms
14
such as algae and bacteria.
Yet there is evidence that flowering plants may have existed during this
time. According to a report in Nature, angiosperm pollen has been found
in Pre-Cambrian rock on the frontier between Venezuela and British
15
Guiana. This rock has been dated by radiometric methods, and has been
16 7
assigned ages of 2,090 m i l l i o n and 1,710 m i l l i o n ' years.
If this report can be taken at face value, it completely upsets the accepted
scientific picture of the origin and evolution of life. If flowering plants were
existing at a time far antedating all known remains of higher organisms,
then the standard evolutionary interpretation of the fossil record must be
mistaken. The obvious implication is that there must have existed Pre-
Cambrian continental regions that were occupied by many different types
of higher organisms, but that were totally destroyed in the course of time.
Such regions can, of course, be compared with the "highlands" posited by
Axelrod to explain the sudden appearance of the angiosperms in the more
recent portion of the fossil record. In this case the evidence directly implies
the existence of such regions, but it still leaves us in the dark as to the mode
of origin of the higher organisms that lived there.
It is perhaps not surprising that evolutionists do not accept evidence that
so blatantly contradicts their theories. Yet the report we have just dis-
cussed is not an isolated instance. There are, in fact, many reports of pol-
len and spores from sedimentary formations antedating the periods in
which the evolution of higher plants is commonly believed to have taken
place. Indeed, there are evolutionists who have accepted some of this evi-
dence and have tried to incorporate it into various modified evolutionary
schemes.
An example of this is provided by a paper by Axelrod that appeared in
18
the journal Evolution in 1959. Therein he cites several reports of spores
and fragments of wood that have been assigned to the Cambrian period.
Some of these spores were found in Kashmir and have been identified as
pteridophytes (ferns), pteridosperms (seed ferns), and gymnosperms (the
19
category of plants including the conifers, or pine trees). Axelrod uses this
and other evidence to completely overturn the standard theory that higher
plants have evolved from Silurian and Devonian psilophytes. He argues
that the different groups of higher plants must all have evolved indepen-
dently from algal ancestors in the Cambrian or earlier. (Axelrod does not
mention, by the way, that angiosperm pollen has also been reported in the
20
Cambrian strata of Kashmir. )
These proposals have apparently not found widespread acceptance, for
current textbooks still expound the psilophyte theory. This becomes espe-
cially interesting when we consider one particular piece of evidence that
Axelrod cites to disprove this theory. According to him, some of the early
Devonian psilophyte fossils are accompanied by fossils of Callixylon logs
21
that are up to three feet in diameter. Callixylon is a type of conifer. One
would think that this fact, if true, would certainly cast doubt on the theory
that all higher plants are descended from these psilophytes.
We can conclude that there is considerable uncertainty anil ambiguity in
our current knowledge of the origin and ancient history of the higher
plants. The fossil record can only be made to support an evolutionary inter-
pretation by the introduction of unverifiable hypotheses. There are also
cases where evidence conflicting with established evolutionary scenarios
seems to have been ignored, or even suppressed. We note that similar
points can also be made about the fossil record of past animal life, although
we do not have enough space to discuss this here.
8.3 The Enigma of Biological Form

If we cannot hope to find direct evidence delineating the development of
major new forms of life, we might at least expect the theory of evolution to
provide us a convincing explanation of how, in principle, such develop-
ments might take place. Since the Darwinian theory asserts that small vari-
ations in organic form gradually accumulate, we might expect evolutionary
theorists to provide us with plausible evolutionary sequences leading from
lower to higher forms of life. In such evolutionary sequences, each or-
ganism should be fit to live in its particular environment, and the differ-
ences between successive organisms in the chain should be of the kind we
would expect from random mutation.
When we examine the literature of evolutionary theory, we do indeed
encounter many explanations of this kind, but in every case they are
disappointingly vague and incomplete. Typical is this statement by the
The Enigma of Biological Form 193
prominent evolutionist Ernst Mayr: "The evolution of the eye ultimately

hinges on one particular property of certain types of protoplasm—photo-
sensitivity. . . . Once one admits that the possession of such photosensi-
22
tivity may have selective value, all else follows by necessity." Mayr does
not, and indeed cannot, specify the particular steps leading from a p h o t o - ,
sensitive speck to a fully developed eye. His account of the evolution of the
eye is typical of theoretical evolutionary explanations, for it relies on an
abiding faith in the power of natural selection and mutation to effect trans-
formations in organic form that evolutionists themselves cannot even"
imagine, much less observe.
Although evolutionists have adhered to this mode of explanation for
many years, there is now evidence that its appeal is beginning to wane. Ac-
cording to a report in Science, the predominant view among the evolutionists
assembled at the recent meeting in Chicago was that the gradual selective
accumulation of small variations cannot account for the appearance of
23
new species.
What has happened is that many evolutionists are now openly ac-
knowledging one of the fundamental problems confronting evolutionary
theory—the problem posed by the complex networks of structure and
function that are characteristic of living organisms. Generally, each com-
ponent of such a network is essential for the proper functioning of the
whole. How, then, could the complete arrangement have arisen through a
finely graded series of functional intermediate forms? In the past many
evolutionists have been content to accept on faith that such sequences of
intermediate stages must always be possible. But now a number of promi-
nent evolutionists are openly admitting that in many significant cases the
required intermediate stages simply may not exist.
To illustrate this problem in evolutionary theory, we shall consider a sim-
ple example provided by a type-of flatworm called the microstomum. 24
This flatworm is equipped with a kind of defensive cell called a nematocyst,

which can fire a poisonous barbed thread. When the flatworm is attacked
by a predator, the nematocysts, situated just beneath the surface of the
worm's back, are discharged, thereby stinging the assailant and driving
it away.
The most interesting aspect of this arrangement is that the nematocysts
are not produced from the tissues of the flatworm itself. Rather, they are
stolen from the hydra, an aquatic organism on which the flatworm preys.
The hydra has tentacles armed with several kinds of nematocysts, which it
uses to subdue and capture the small animals on which it feeds. Some of
these cells fire poisonous barbs, and others discharge various types of
coiled and sticky threads that enable the hydra to hold on to its prey.
I'M The Doctrine of Evolution
These flatworms generally avoid hydras. But biologists have observed

that when one of them needs more nematocysts, it will eat hydras and di-
gest all of their tissues except these particular cells. The nematocysts are
neither damaged nor discharged, but are enclosed within certain of the flat-
worm's cells, which carry them toward the worm's back. The nematocysts
that fire coiled or sticky threads are then digested, but those that fire
poisonous barbs are transported to sites just beneath the outer layer of the
worm's back.
There the nematocysts are oriented so that their stings will fire upward.
The epithelial cells which form the worm's outer layer become very thin
just above the newly positioned nematocysts, thus providing portholes for
the firing of the stings. Finally, the cells that have encapsulated the nemato-
cysts undergo extensive changes that enable these cells to act as trigger
mechanisms. (The hydra's original trigger mechanism is contained in a type
of cell called a cnidoblast, which the flatworm digests.)
Let us consider whether or not these defensive arrangements of the
microstomum could have evolved step by step. An evolutionary scenario
would have to begin with an ancestral flatworm that ate hydras but did not
make use of their nematocysts. In such a worm, what would be the first
evolutionary step leading to the eventual exploitation of the nematocysts
as defensive weapons? Unless the nematocysts were actually used as
weapons, for the worm to manipulate them internally would be useless.
Indeed, it would be dangerous, since the flatworm can easily be killed by
the discharge of the hydra's stings.
Yet each step in the internal processing of the nematocysts is essential for
their eventual use as weapons. If they were not transported to the flat-
worm's back, they could not be usefully deployed. If transported to the
back but oriented incorrectly, they would be useless or even dangerous.
If they were oriented beneath epithelial cells of normal thickness, the
discharged sting would lose its momentum while passing through the
epithelium, and the worm would sting itself.
There is also a further problem. Evidently the nematocysts are not
triggered simply by pressure applied to the worm's back. Rather, their
firing is governed by a complex control mechanism within the worm. With-
out this trigger mechanism the whole arrangement would be useless, even
if the nematocysts were properly oriented beneath epithelial "portholes."
When examined closely, each step in the internal manipulation of the
nematocysts resolves into a complex of substeps. For example, for a
nematocyst to be transported to the back of the flatworm, one of the
worm's cells must first recognize it, and then the cell must initiate a process
The Enigma of Biological Form 195
of motion that specifically carries the nematocyst to the worm's dorsal re-
gion. These are both complex procedures. Yet for the flatworm to take ad-
vantage of the hydra's nematocysts, it would seem that many complex
arrangements of this kind must be present simultaneously.
It is hard to see how complicated, interlocking arrangements such as the
microstomum's defensive system could have evolved by many small steps
in the traditional Darwinian manner. Darwin himself maintained, " I f it
could be demonstrated that any complex organ existed which could not
possibly have been formed by numerous, successive, slight modifications,
25
my theory would absolutely break d o w n . " Yet what possible sequence of
intermediate forms could span the gap between an ordinary flatworm and
a flatworm capable of deploying stolen nematocysts for its own defense?
It would seem that we are confronted with one of two alternatives: either
(1) some of these hypothetical intermediate forms must have useless or
deleterious features, or (2) some of these forms must be separated by large
gaps corresponding to many simultaneous modifications. It is quite pos- *
sible that a series of intermediate forms of the kind required by Darwin
simply does not exist.
Many organs are so complex that we do not clearly understand how
they work. One might think that evolutionists would refrain from offering
evolutionary explanations for such organs. Unfortunately, it is precisely
in these cases, where our understanding is vague and incomplete, that
evolutionists seem most eager to explain origins by waving the wand of
Darwinian natural selection. For example, one prominent exponent of the
evolutionary viewpoint has argued in a popular presentation that a mere in-
26
crease in size can transform an ape's brain into a human b r a i n . He claims
this could be accomplished by the natural selection of a series of slight pro-
longations of the period of foetal brain growth. But can a mere increase in
brain size account for the difference in mentality between human beings
and apes? It is possible to entertain such fantasies only because we are now
almost entirely ignorant of how the brain works.
Simpler organs are easier to analyze, and we can frequently see, as we
did in the case of microstomum, that they are not amenable to explanation
by traditional evolutionary concepts. To illustrate this, we will briefly
consider three additional examples of simple organ systems.
Our first example is provided by the statocyst of a certain species of
27
shrimp. The statocyst is a hollow, fluid-filled sphere built into the
shrimp's shell. It is lined with cells bearing pressure-sensitive hairs, and it
contains a small weight. The weight tends to sink and press against the
downward portion of the sphere, thus enabling the shrimp to tell up from
down. Curiously, the weight is a small grain of sand that the shrimp picks
up with its claws and inserts into the statocyst through a small hole in its
shell. Since the statocyst is built into the shell, the shrimp has to do this
every time it moults.
Now, by what intermediate stages could the shrimp's statocyst have
arisen? Both the statocyst and the behavioral pattern involved in picking
up the grain of sand are quite complex, and neither is of any use without the
other. Even if the statocyst had evolved with a built-in weight and then had
lost this feature by a mutation, the appearance of the insertion behavior
would require a leap involving the coordination of many variables.
For our second example we turn to the bacterium Escherichia coli, an
organism that is thought to be very low on the evolutionary scale. Each
E. coli bacterium possesses several long, helically curved fibers (called
28
flagella) that enable it to s w i m . Each flagellum is connected at one end to
a kind of motor built into the bacterial cell wall. When these motors rotate
in a certain direction, the flagella rotate in unison and act as propellers to
drive the bacterium forward through the water. When the motors rotate in
the opposite direction, the flagella separate and change the orientation of
the bacterium by pulling in various ways. By systematically alternating be-
tween these two modes of operation, the bacterium is able to swim from
undesirable to desirable regions of its environment.
Some investigators have proposed that the motors are driven by a flux of
29
protons flowing into the c e l l . According to them, each motor consists of
a ring of sixteen protein molecules attached to an axle and a stationary ring
of sixteen proteins built into the cell wall. Protons are steadily pumped out
of the cell by its normal metabolic processes. As some of these protons flow
back into the cell through the pairs of rings, they impart a rotary motion to
the movable ring. Since the motor can operate in forward or reverse, there
must be some mechanism that adjusts the configuration of the molecules in
the rings so as to reverse the torque produced by the flow of ions.
Although the exact details of the E. coli's molecular motors have not yet
been worked out, we can see that their operation depends on the precise,
simultaneous adjustment of many variables. Even though these motors are
fairly simple in structure, it is very hard to imagine a continuum of useful
forms spanning the gap between a motorless cell and a cell with a fully func-
tional motor. It is quite possible that no such continuum of forms exists, or
that such a continuum must involve intermediate structures much more
complicated than the existing molecular motors. In any event, we presently
have no idea how the E. coli's motors could have evolved. Indeed, the
problem becomes even more difficult when we consider that these motors
The Ressurrection of the Hopeful Monster 197
are useless without their control systems, and the control systems are
useless without the motors.
For our final example we consider a kind of marine invertebrate called a
sea slug. Certain of these sea slugs are able to steal the nematocysts from
sea anemones (a marine creature similar to the hydra) and use them for
3
their own defense. " Their system is quite similar to that of microstomum,
except that instead of transporting the nematocysts by special migratory
cells, the sea slugs move them into position by sweeping them through
narrow ciliated tubes that pass from their stomachs to their backs.
Now, the sea slug is a mollusc, and therefore cannot be closely related
to the flatworm, microstomum. Evolutionists customarily explain the exis-
tence of similar traits in ostensibly unrelated organisms by invoking the
idea of convergent evolution. According to this idea, the forces of evolu-
tion will automatically produce similar effects in similar circumstances. But
the problem we face here is that by the hypothesis of gradual evolutionary
change, we cannot explain the origin of the defensive system of either the
sea slug or microstomum.
8.4 The Resurrection of the Hopeful Monster

Clearly we must reject the standard evolutionary theory of gradual
change through the natural selection of many small variations. How, then,
can the origin of species be explained? At the conference in Chicago, there
were signs that at least some evolutionists are trying to resurrect a theory
that was greeted with almost universal scorn and derision when first pro-
31
posed in the 1940s. This is the theory of the "hopeful monster," devised
by the geneticist Richard Goldschmidt.
The key to this theory is the idea that the genetic systems of organisms
must be so arranged that a single mutation can produce, in one stroke, an
elaborate systematic change in biological structure and function. Almost
all known mutations result in gross defects, and a few result in small modifi-
cations that can be useful for the organism under suitable environmental
conditions. According to Goldschmidt, however, there must exist a special
type of mutation capable of generating new complex structures, such as
functional legs, wings, or lungs. Most of these macromutations, as he called
them, would result in bizarre monstrosities completely unfit for survival.
But a few macromutations would produce "hopeful monsters," novel
creatures that just happened to be adapted to a totally new mode of
32
existence.
This theory is now at a very tentative and speculative stage, and many
evolutionists still view it with suspicion. Nonetheless, it represents an im-

portant trend in current evolutionary thought, and it illustrates the desper-
ate extremes to which evolutionists have been forced to go in their efforts
to construct a workable evolutionary theory. We shall therefore briefly
consider some of the reasoning underlying the theory of the hopeful
monster.
In its present recension, this theory relies on the concepts of regulative
33
and structural genes. Biologists define a structural gene as a sequence of
D N A coding that defines a specific structural element of a living organism.
An example is the gene for hemoglobin, the oxygen-carrying pigment of the
red blood cells. In contrast, a regulative gene is a sequence of D N A coding
that controls the timing and order of expression of other genes. We can en-
vision an interacting system of regulative and structural genes that acts as a
kind of genetic computer program. Such a program might be capable of ex-
pressing and repressing various combinations of structural genes in a com-
plex and systematic way. The hopeful monster theory proposes that small
changes in such genetic programs might result in the systematic large-scale
changes in biological organization that evolutionary theorists need.
Biologists cite certain kinds of mutations as evidence for the existence of
regulatory genes. For example, sometimes horses are born with three-toed
feet. One might tentatively explain this variation by saying that although
the genetic system of the horse always has the structural information for a
multitoed foot, in normal horses a regulatory gene suppresses the genes for
all the toes but one. When a mutation disables this regulatory gene, the
latent genetic information is expressed, and a multitoed horse is born.
A certain mutation of fruit flies provides another possible example of the
interaction of regulative and structural genes. In this mutation, known as
aristapedia, a fully developed leg grows from the head of the fly in the posi-
tion where the antenna normally grows. Scientists explain this anomaly by
proposing that the regulative and structural genes for the leg comprise a
kind of "subroutine" that is set in motion under the control of other regula-
tory genes. These regulatory genes may store information specifying the
location of the leg, and if this information is disrupted by a mutation, the
leg may form in an abnormal place.
Although the theory of regulative genes is still highly speculative, it does
not seem unreasonable as a way of explaining certain types of mutation.
But how this theory can account for Goldschmidt's hypothetical "macro-
mutations" that produce complex, finely coordinated organs in a single
stroke is not at all clear.
We shall try to illustrate the potentialities and limitations of systems of
regulatory genes by constructing a simple artificial example. We can regard

the following array of symbols as a "genetic" system for a series of English
sentences.
I am in (2); the orthodox (1) in (3). I believe I am in (9) as an (8) would be in if

set to learn (5). The (8) (1) it was (7); ( ) yet I cannot keep out of (4).
1—would say; 2—thick mud; 3—fetid abominable mud; 4—the question;
5—the first book of Euclid; 6—and 1 am in much the same mind; 7—no man-
ner of use; 8—old gorilla; 9—much the same frame of mind.
The code used in this genetic system is almost self-explanatory. To produce

the encoded English statements, simply replace each number in paren-
theses with the corresponding numbered phrase. The result is the following
statement made by Charles Darwin about the mechanism of large-scale
evolution.
I am in thick mud; the orthodox would say in fetid abominable mud. I believe I
am in much the same frame of mind as an old gorilla would be in if set to learn
the first book of Euclid. The old gorilla would say it was no manner of use; yet
14
I cannot keep out of the question.
In our artificial genetic system, the numbers in parentheses play the role
of regulatory genes, and the phrases play the role of structural genes. If we
mutate the regulatory gene (4), changing it to (3), we shall observe a
change in Darwin's statements similar to the aristapedia mutation of fruit
flies. (We invite the reader to try this and observe the effects.) Also, when
we examine the genetic system closely, we find that a mutation has changed
one regulatory gene to ( ). If we convert this gene to (6), Darwin's state-
ment apparently acquires an entirely new sentence, although all that has
actually happened is that the complete text of the original has been
restored.
We can thus see that various kinds of large-scale effects result from
mutations in the regulatory genes of our artificial system. Yet all of these
effects have one thing in common. They all involve the manipulation of ma-
terial already present in the genetic system. To induce the system to
produce something entirely new is a different matter.
For example, we invite the reader to try to find mutations that will ex-
pand Darwin's remarks to include the following statement from his Origin
of Species.
I can see no difficulty in a race of bears being rendered, by natural selection,

more and more aquatic in their habits, with larger and larger mouths, till a
1
creature was produced as monstrous as a whale. ''
We find ourselves in a quandary. We can gradually build up this additional

statement by many small mutations, each of which might occur by "chance"
with a reasonably high probability. But the intermediate states will all en-
tail nonsensical sentence fragments that correspond in our analogy to
useless or harmful mutant organs. We can also introduce the entire state-
ment by a single random mutation; but then we are confronted by the prob-
lem that such a mutation must be exceedingly improbable. The more
, numerous the letters involved, the more improbable it is that they will
fall into place the way you want them. The probability goes down exponen-
tially with the number of variables, and the same can be said in general
about multifeatured biological mutations.
Of course, we could devise a genetic system in which a mutation in a
single regulatory gene would cause our new statement suddenly to mani-
fest itself. But could we do this without, in effect, building the statement
into the system? If not, one might naturally ask how such complex latent
information got into the genetic system in the first place.
Questions such as these cannot be avoided in the study of bears, whales,
and living beings in general. The point of our artificial example is that the
concepts of regulative and structural genes, although suggesting possible
ways to explain several kinds of mutations, do not automatically answer
these questions. The real problem of evolutionary theory—how to account
for the origin of completely new organs and functions—remains as baffling
as before. A n d until evolutionists can provide a convincing solution to this
problem, we must conclude that their evolutionary speculations have no
sound basis.
As a final observation, we note that the concept of regulatory genes can
be used as the starting point of a nonevolutionary theory of the origin of
species. Consider the datum that horses are occasionally born with three-
toed feet. One explanation of this is that horses are the lineal descendants
of multitoed ancestors. According to the latest theories, in the course of the
horses's evolutionary history, a series of mutations in regulatory genes
must have suppressed the embryonic development of all but the central
toe.
An alternative explanation runs as follows: Let us suppose that an intelli-
gent creator wished to devise blueprints for a variety of animal forms. What
would be the most economical way for him to go about this? Would it be
best for him to design each animal completely from scratch? Or would it be
more efficient for him to devise one basic plan that could be modified in
various ways to produce various specific animal forms?
We can see that the second strategy would be the most economical, and it
is certainly the strategy that a human engineer would choose, if possible.

Now, it turns out that the idea of regulative genes provides a means of
executing this strategy. Suppose, for example, that a highly intelligent en-
gineer had decided to design the mammals by making modifications in a
fundamental five-toed body plan. In that case, when he began to work on
the design for a horse, he would be faced with the problem of how to elimi-
nate all but one of the toes of the generalized mammalian foot. Our propo-
sal is that this could easily be done by the manipulation of regulatory genes
if the engineer had had the foresight to properly design the genetic system
of his basic mammalian plan. If the genetic system were properly designed,
he could suppress the growth of the unwanted toes simply by throwing a
single genetic switch, and thus save himself the effort of completely
redesigning the genes for the foot.
We should point out, however, that while such manipulations of the
genetic system might simplify the task of designing a horse's foot, they
could not be sufficient for the complete execution of this task. Consider the
following description of a horse's hoof:
Such a hoof, which is fitted to the limb like a die protecting the third phalanx,
can without rubber or springs buffer impacts which sometimes exceed one
ton. It could not have formed by mere chance: a close examination of the
structure of the hoof reveals that it is a storehouse of coaptations and of or-
ganic novelties. The horny wall, by it's vertical keratophyl laminae, is fused
with the podophyl laminae of the keratogenous layer. The respective lengths
of the bones, their mode of articulation, the curves and shapes of the articular
surfaces, the structure of bones (orientation, arrangement of the bony layers),
the presence of ligaments, tendons sliding with sheaths, buffer cushions,
navicular bone, synovial membranes with their serous lubricating liquid,
all imply a continuity in the construction which random events, necessarily
36
chaotic and incomplete, could not have produced and maintained.
The author of this description is observing that the horse's hoof has many
complex, intricately interrelated features that can hardly be accounted for
by random mutations in either structural or regulatory genes. To design a
regulative gene system that would allow for the quick production of such an
elaborate structure, much work would be required—perhaps as much
work as would be needed to design the hoof completely from scratch. Here
our engineer would have no recourse but to directly confront a difficult
problem in complex design. This, of course, is possible for an intelligent en-
gineer, but no one has yet shown how it could be accomplished by a blind
evolutionary process.
The baleen whales provide another example of how regulative genes could
be used to facilitate the production of organic designs. These whales do not

have teeth, and their mouths are equipped with an elaborate filter that they
use to strain from the seawater the small organisms on which they feed. The
embryo of a baleen whale begins to develop teeth, but these are resorbed
by the tissues of the mouth at an early stage of development. Evolutionists
have argued that these embryonic teeth prove that the baleen whales must
37
have descended from a toothed ancestor.
Yet a different interpretation is possible. Let us suppose, as before, that
an intelligent engineer was designing the forms of various mammals by
modifying a basic mammalian plan. Such a plan would naturally have speci-
fications for teeth, but if the plan were properly designed, the engineer
could suppress the development of the teeth by making simple adjustments
in regulatory genes. Such adjustments would not eliminate all of the speci-
fications for teeth from the genetic system, and they might allow for the
temporary appearance of teeth in the embryonic stage, as we see in the
whale. To produce the baleen filter, however, the engineer would have to
introduce many specific, detailed instructions into the genetic system.
We can thus see that the supposedly atavistic teeth of the embryonic
whale can also be given a nonevolutionary interpretation as byproducts of
the design strategy of an intelligent engineer. The whale's filtration system
can also readily be understood as a product of intelligent design, but it is
hard to see how it could have evolved by mutation and natural selection.
The hypothesis that living organisms have been designed by an intelli-
gent creator is in agreement with the alternative world view that we have
been introducing step-by-step in this book. We cannot conclude, of course,
that the structural plans of living organisms were necessarily conceived and
executed in the particular way that we have envisioned in this speculative
discussion. Our aim was to show that evidence commonly cited as proof
of the theory of evolution can often be explained just as convincingly in
nonevolutionary terms. However, such speculative explanations cannot
actually provide reliable information about the process of creation or the
nature of the creator. Our theme has been that such information can only
be obtained by a transcendental process that takes advantage of the higher
sensory and cognitive capacities of the conscious self.
8.5 Evolution and Negative Theology

We have seen that there is no direct evidence for the evolution of com-
plex organic forms and that some prominent paleontologists maintain that
such evidence may never be found. We have also seen that evolutionists do
not have an adequate theory of evolutionary change and that they are still
Evolution and Negative Theology 203
groping for such a theory in the realms of speculation and vague conjec-
ture. We are therefore led to ask: In the absence of both observation and
theory, what has convinced scientists to accept what we can only call the
doctrine of evolution?
One important line of reasoning that has led many persons to adopt an
evolutionary point of view could be called the argument by negative theol-
ogy. Darwin himself used this argument extensively, and since his time it
has been a mainstay of evolutionary thought.
In a recent popular book, the paleontologist Steven J. Gould presents
one form of the negative theological argument in these words: " O d d ar-
rangements and funny solutions are the proof of evolution—paths that a
sensible God would never tread but that a natural process, constrained by
38
history, follows perforce." The general form of the argument can be out-
lined as follows: " G o d must have certain characteristics, X, and therefore
v
He would have created a certain sort of world. Since the world as we see it i s
very different from this, it must be that there is no God. Since the only
alternative to divine creation that we can think of is evolution, life must «
have arisen by some kind of evolutionary process."
This argument has two basic forms. One of these is the traditional ar-
gument from evil against the existence of God. According to this argu-
ment, the existence of many kinds of suffering, both in the human species
and in the plant and animal kingdoms, is inconsistent with the idea that
the world was created by an all-powerful, benevolent being. In contrast,
such suffering seems to fit naturally into the evolutionary world view. %
The second form of the argument is that many features of living organ-
isms would not, as Gould says, be designed by a "sensible G o d " and must
therefore be due to evolution. An example of this sort of argument is pro-
vided by Darwin's work with orchids. Darwin observed that the petals of
these flowers are deployed in many remarkable arrangements, which in-
sure that visiting insects will carry pollen from one flower to another. Yet
since modified petals, rather than a completely novel kind of structure, are
used in these arrangements, Darwin argued that divine creation was ruled
out and that the orchids must therefore be products of evolution. In the
words of Gould, " I f God had designed a beautiful machine to reflect his
wisdom and power, surely he would not have used a collection of parts
39
generally fashioned for other purposes."
What can we say about these arguments? We can immediately dismiss
both versions of the negative theological argument as scientifically un-
sound, for they are based on completely speculative ideas about the pur-
poses of God and the methods He uses to achieve them. We have pointed
out before that while reasoning that rests on a finite set of material observa
tions may confirm a given theory about God, such reasoning cannot prove
the theory to be true. The same observation can be made about negative
arguments that depend on a particular theological theory. In the negative
theological argument the initial premise, "God must have certain charac
teristics, X," has never even been clearly formulated by the evolutionists,
and such a premise has certainly never been proven by them, either by logic
or by induction from observations. It is also hardly necessary to point out
that the second premise of this argument—the premise that evolution is the
only alternative to divine creation—has also never been proven.
The negative theological argument is dependent on a painfully naive and
limited conception of God, and it collapses as soon as a more satisfactory
conception is introduced. We have already seen an example of this in our
discussion of the horse's toes and the embryonic teeth of the whale.
Evolutionists argue that a "sensible G o d " would never produce such aber
rations, but we have seen that these bodily structures can readily be inter
preted as byproducts of the design strategy of a master engineer. The key to
this interpretation is the realization that God does not necessarily intend
the material world as an exhibition of His finest worksmanship.
What is the purpose of the material world? The evolutionists have
adopted a mechanistic world view that excludes the very idea of purpose,
but at the same time some of them cling to a certain conception of what
God's purposes would be if He existed. Yet theirs is not the only conception
that is possible. To demonstrate this, we will outline the understanding of
the purpose of the material world presented in the Bhagavad-gītā.
According to the Bhagavad-gītā, one can understand the purpose of the
material creation through the concept of free will. The natural relationship
between the jīvātmā and Krsna is one of loving reciprocal service, and love
can exist only for one who has freedom. The jīvātmā is free to turn away
from his relationship with the Supreme Person and seek to be independent,
and Krsna creates the material world as a place where the jīvātmā can do
this. Here the jīvātmā becomes temporarily forgetful of his true nature and
40
transmigrates from body to body in various species of life.
Therefore, the material world is a place of suffering. The conditioned
living beings, deprived of their central object of devotion, inevitably have
clashing interests, and become intense sources of misery for one another.
This misery is a byproduct of the more fundamental misery of ignorance
and forgetfulness that characterizes the material world, and is exhibited in
various degrees of severity in various forms of embodied life.
The bodies of the living species are temporary vehicles designed to ac
commodate sentient beings in various states of forgetful consciousness.
Evolution and Negative Theology 205
Since all of these forms represent limitations on the true nature of the
jīvātmā, it is not surprising that they should be crafted in a rough and ready
manner. It is also not surprising that the living species should show sys
tematic similarities in bodily structure. A l l organic forms correspond to
states of consciousness of one fundamental type of being, and, in principle,
they can all be understood systematically in psychological terms.
According to the Bhagavad-gītā, the miserable conditions in the ma
terial world are due neither to malevolence nor to poor design on the part of
the creator. Rather, they are consequences of the free will of the jīvātmā
himself, and they can be alleviated on an individual basis by the proper
exercise of free will. In Chapter 9 we will discuss this in greater detail. Here
we simply note that the negative theological argument of the evolutionists
does not apply to the world system of the Bhagavad-gītā—at least, not as
the argument stands. Evolutionists would do well to examine the hidden
metaphysical assumptions of their argument, and consider whether or not
such an argument can serve as the basis for a scientific theory.
If we closely examine the negative theological argument, we find that it
seems to originate not from logical considerations, but from a sense of deep
dissatisfaction with the theological conceptions that form its real founda
tion. These conceptions are inherited from Western religious systems that
the evolutionists, beginning with Darwin, have emphatically rejected. Pre
sentations of the theory of evolution are often permeated with an attitude
of contempt for these religious systems, and the negative theological argu
ment, in particular, often takes the form of an emotional tirade directed
against the idea of divine creation.
When we consider the irrational character of such presentations in light
of the observational and theoretical weaknesses of evolutionary thought,
the "theory" of evolution seems little more than a poorly reasoned intellec
tual reaction against a spiritual tradition that was perceived as inadequate.
Unfortunately, it has also been an entirely futile reaction, for the evolu
tionists have succeeded neither in providing a genuine alternative source of
spiritual knowledge nor in establishing a workable material explanation for
the origin of life.
The role the negative theological argument plays in the theory of evolu
tion becomes clear when we consider the historical context in which this
theory arose. When Darwin published his Origin of Species in 1859, scien
tific thought in Europe had been dominated for many years by an approach
to spiritual knowledge known as natural theology. According to this ap- -
proach, one can deduce from observations of natural phenomena that the
world has been created by a supremely intelligent, benevolent, and all- ,
powerful being. Pointing to the highly organized plans of living beings, the
The Doctrine of Evolution
proponents of natural theology evoked the "argument from d e s i g n w h i c h

maintains that these plans imply the existence of an intelligent creator.
This emphasis on natural theology indicated a serious weakness in the
philosophical and religious thought of the time. The argument from design
was stressed because it seemed to provide definite proof of the existence
of God. Yet this argument has two deficiencies—it is not able to actually
prove that God exists, and more importantly, it is not able to provide
definite information about the nature of God.
The conception of God advocated by the natural theologians was actu-
ally derived from the traditional religious systems of Judaism and Chris-
tianity. The verifiability of this conception was entirely dependent on the
capacity of these religious systems to bring individuals into direct contact*
with God. The argument from design could only provide confirmation
and the illusion of verification, and the fact that it was heavily stressed
shows that practical means of attaining genuine spiritual realization were
wanting.
To many scientists this state of affairs was disheartening. The argument
from design seemed to lock them into a stultifying system of useless, un-
verifiable doctrines. It could not satisfy their curiosity about the origins of
life, it offered no new avenues of scientific investigation, and it provided no
tangible spiritual insight. Thus many perceived natural theology as a dead
end.
In this context we can understand the initial appeal of Darwin's theory
of evolution. Darwin explicitly rejected the sterile arguments of natural
theology and introduced an approach to the origin of life based entirely on
physical principles. This approach seemed to bring the question of origins
into the familiar realms of physics and chemistry, where scientists had had
so much success using the experimental method, and thus it held forth the
promise of similar success.
Yet we have seen that this promise has never been fulfilled. Although the
theory of evolution has become institutionalized as the standard explana-
tion for the origin of species, it has not outgrown its original status as a reac-
tion against an imperfect and restrictive system of theological thought. As
such, it is no more valid than the system of natural theology it historically
displaced. If speculative reasoning has given rise to an imperfect concep-
tion of G o d , then we certainly cannot expect negative arguments based on
this conception to yield a true understanding of life's origins.
H o w , then, can we obtain a genuine understanding? Our suggestion is
that we can reach such an understanding only through a valid spiritual sci-
ence. Speculation resting on a finite set of material observations is indeed
inadequate to provide valid knowledge about a supreme transcendental
Notes 207
being. But the answer to this problem is not to deny the existence of such a
being and to seek explanations solely in familiar physical principles. This is
the fallacy of the drunk who lost his keys near the doorstep of his house but
would search for them only under a streetlamp because the light was better
there.
Notes
1. Tax and Callender, eds., Evolution After Darwin, pp. 265-266.
2. Rensberger, " R e c e n t Studies Spark Revolution in Interpretation of

Evolution," p . C 3 .
3. Ibid.
4. Gould and Eldridge, "Punctuated Equilibria: T h e T e m p o and M o d e of
Evolution Reconsidered," p p . 115-151.
5. Grasse, Evolution of Living Organisms, p. 124.
6. H a r l a n d , et. al. (eds.), The Fossil Record.
7. Ibid.
8. A n d r e w s , Studies in Paleobotany, chap. 6.
9. Axelrod, "The Evolution of Flowering Plants," p. 229.

10. Axelrod, p. 230.
11. Ibid.
12. Dodson and D o d s o n , Evolution, Process and Product, p. 141.

13. Valentine, "The Evolution of Multicellular Plants and Animals,"
p p . 141-158.
14. Glaessner, "Biological Events and the Precambrian Time Scale,"

pp.470-477.
15. Stainforth, pp. 292-294.
16. McDougall, C o m p s t o n , and H a w k e s , p. 567.
17. Snelling, p. 1079.

18. Axelrod, "Evolution of the Psilophyte Paleoflora," p p . 264-275.
19. J a c o b , J a c o b , and Shrivastava, p. 166.

20. Wadia, The Geology of India, p p . 141-142.
The Doctrine of Evolution
21. Axelrod, "Evolution of the Psilophyte Paleoflora," p p . 268-269.
22. Mayr, " T h e E m e r g e n c e of Evolutionary Novelties," p. 359.
23. Lewin, "Evolutionary T h e o r y U n d e r F i r e , " p. 883.
24. K e p n e r , Gregory, and Porter, " T h e Manipulation of the Nematocysts

of Chlorohydra by M i c r o s t o m u m , " p p . 114-124.
25. D a r w i n , On the Origin of Species, p. 189.
26. G o u l d , The Panda's Thumb, p. 133.
27. B u d d e n b r o c k , The Senses, p p . 138-141.
28. Berg, " H o w Bacteria Swim," p p . 36-44.
29. Hinkel and McCarty, " H o w Cells M a k e A T P , " p. 116.
30. Wilson, The Mystery of Physical Life, pp. 28-31.
31. Adler, "Is M a n a Subtle A c c i d e n t ? " p p . 95-96.
32. Goldschmidt, The Material Basis of Evolution.
33. G o u l d , " H e n s T e e t h and H o r s e ' s T o e s , " p p . 2 4 - 2 8 .
34. Cited in Gillespie, Charles Darwin and the Problem of Creation,

p . 87.
35. D a r w i n , p. 184.
36. G r a s s e , p. 5 1 .
37. G o u l d , The Panda's Thumb, p. 29.
38. Ibid., p p . 2 0 - 2 1 .
39. Ibid.
40. A . C . Bhaktivedanta Swami Prabhupāda, Bhagavad-gītā As It Is,

p p . 8 9 , 1 0 2 , 704-707.
PART III
CONCLUSION
Chapter 9
The Epistemology
Of Transcendental
Consciousness
In a letter to Max B o r n , Albert Einstein expressed his faith as a physicist:
You believe in the God who plays dice, and I in complete law and order in
a world which objectively exists, and which I, in a wildly speculative way,
am trying to capture. I firmly believe, but T hope that someone will discover
a more realistic way, or rather a more tangible basis than it has been my lot
to do.'
Einstein believed that the phenomena of nature were expressions of a uni-

fied underlying reality, and he was convinced that this ultimate foundation
must be accessible to human understanding. Following the path of modern
science, he believed that the fundamental principles of nature could be ex-
pressed by mathematical formulas abstracted from experimental data by
the power of his mind. According to Einstein, scientists could pin down
the unified basis of reality if they could find simple formulas that would
account for all the phenomena of the universe.
Yet we have seen in Part II that a quantitative description of nature can-
not be both simple and universal. In particular, we have shown that any
quantitative description that adequately accounts for the phenomena of
life must be exceedingly complex. When a large amount of information is
expressed as succinctly as possible, -it satisfies the statistical criteria for ran-
domness; thus a truly complex law, if it is reduced to its simplest f o r m , will
appear to be a product of lawless chaos. (See Chapters 5 and 6.) It follows
that Einstein's method of capturing the world can end only in frustration.
In the search for a unified understanding of nature, we must ultimately
confront a vast array of irreducibly complex patterns that cannot be speci-
fied by a simple universal law or formula, and that must therefore seem
extremely arbitrary from a mechanistic point of view.
We encounter such irreducible patterns when we try to explain living
form in its various aspects—which range from the metabolic processes of
cells to the subtleties of human behavior. In an ultimate sense, such pat-
terns are inexplicable, for any quantitative explanation that is intended
212 The Epistemology of Transcendental Consciousness
to account for them must be just as complex and arbitrary as the patterns
themselves.
While the complex bodily structures of living organisms cannot be con
cisely explained in mechanistic terms, at least they can be quantitatively
described. In Part I of this book, however, we discussed another feature of
life—the phenomenon of conscious awareness—that is not even touched
upon by descriptions of the measurable behavior of matter.
We customarily associate consciousness with certain states of activity in
"conscious living bodies," but since we directly perceive our own con
sciousness, consciousness must be more than just a name for particular pat
terns of behavior. One may say that this perception is subjective, but if we
are to avoid the position of solipsism, we must accept the consciousness of
others to be an objective fact of nature. This objectively existing conscious
ness must be distinguished from the behavior of matter. Indeed, if we
analyze the functioning of material systems such as the brain, we see that
the contents of our consciousness cannot be correlated in a one-to-one
fashion with measurable events in such systems. (See Chapters 2 and 3.)
Consciousness is thus a feature of reality that is impossible to capture by
mechanistic laws, either simple or complex.
The phenomena of consciousness and complex form stand as insurmount
able obstacles blocking any attempt to capture the world by a quantitative
theory. To find a successful approach to understanding reality, we must
therefore depart from the mechanistic framework of modern science. In
this book we have taken a step in this direction by outlining an alternative,
nonmechanistic world view based on the Bhagavad-gītā, the Bhāgavata
Purāna, and other Vedic literatures of India.
The world view of the Bhagavad-gītā is based on the postulate that con
scious personality is the ultimate basis of reality. Thus far we have intro
duced some of the elements of this world view by briefly describing two
fundamental categories of conscious beings. The first category has a sin
gle member—the unique Supreme Person, Krsna, who is the primordial
cause of all causes, and who is directly conscious of all phenomena. The
second category consists of the innumerable localized conscious beings, or
jīvātmās. The jīvātmās are irreducible conscious persons, and they are
qualitatively the same as the Supreme Person. Yet they differ from the
Supreme in that they are minute and dependent, whereas the Supreme
Person is unlimited and fully independent.
The nature of the jīvātmās is discussed in Chapters 1, 2, and 3, and the
Supreme Person is described in Chapter 6. Chapter 7 discusses the rela
tionship between the Supreme Person, the jīvātmās, and the material
The Epistemology of Transcendental Consciousness 213
energy, and it introduces the solution to the mind-body problem given in

the Bhagavad-gītā.
We argue in these chapters that the philosophy of the Bhagavad-gītā
provides a consistent picture of the phenomena of life. It accounts for the
origin and continued maintenance of the complex material forms of living
organisms. It also clarifies the nature of individual consciousness, and
explains the relationship between the conscious self and the body. The
objection may be raised, however, that even though this philosophy may
provide interesting speculative solutions to certain fundamental scientific
problems, we have not yet even begun to prove that it is true.
We agree that nothing we have said so far constitutes proof. The two
categories of conscious beings that we have considered lie almost entirely
outside the purview of empirical investigation based on reason and ordi
nary sense perception. Our conscious awareness includes direct perception
of itself, but apart from this, our ordinary senses provide us only with in
formation about the configurations of material bodies. On the basis of
reasoning, introspection, and ordinary sense perception, we can infer that
consciousness must be due to some entity that is distinct from matter as we
know it, but we cannot arrive at a truly satisfactory understanding of this
entity by these means.
Similar remarks can be made about the problem of proving the existence
of a supreme conscious being. Many philosophers and scientists have main
tained that the complex order displayed by the bodies of living organisms is
evidence for the existence of an intelligent creator. We have argued that
this is indeed a reasonable explanation of biological f o r m , and we have
shown in Chapters 5 through 8 that scientists of the evolutionary persua
sion are still groping unsuccessfully to find a workable mechanistic expla
nation. Yet observations of biological form convey no clear picture of the
creator by themselves, and it is indeed hard to see how a finite number of
observations made within a limited region of space anil lime could prove
very much about the nature of an unlimited eternal being.
Arguments for the existence of God from the evidence of nature are usu
ally based indirectly on a conception of God that is derived from other
sources. These arguments may show that such a conception is consistent
with the facts of nature, but what these facts actually entail is at best an idea
of God so vague and general that it is practically useless.
So, if our alternative model of reality cannot be established by standard
empirical methods, then how can it be established? We have suggested that
the key to verifying our model is provided by the unique nonmechanistic
features of the model itself. According to the Bhagavad-gītā, the natural
senses of the jīvātmā are not limited to picking up information from the
sensory apparatus of a particular material body. Indeed, a jīvātmā in this
situation is considered to be in an abnormal condition, and can be com
pared to a person who has become so engrossed in watching a television
program that he has forgotten about his own existence, and has accepted
the flickering two-dimensional image on the television screen to be the all
in all. The jīvātmā is capable of directly perceiving both other jīvātmās and
the Supreme Person, although in the bodily state of existence, he is pre
vented from doing this by his preoccupation with the fascinating show
presented by the bodily senses.
It follows that our model can be verified if a way can be found to re
awaken the full cognitive capacities of the conscious self. In this chapter we
will outline a practical method of doing this known as bhakti-yoga, or devo
tional service. We will present this process as a method of obtaining reliable
knowledge about aspects of reality that are inaccessible by traditional
methods of scientific research. We should note, however, that bhakti-yoga
is not simply a method of obtaining knowledge. Rather, it is a means
whereby each individual conscious self can attain the ultimate goal of his
existence.
9.1 The Process of Bhakti-yoga
The process of bhakti-yoga involves reawakening the natural relation

ship that exists between the individual conscious self and the Supreme
Person, Krsna. We pointed out in Chapter 7 that Krsna accompanies each
embodied jīvātmā as the Supersoul, and that He directs the material body
of the jīvātmā in accordance with the jīvātmā's desires and past karma. This
means that a relationship is always existing between the jīvātmā and the
Supreme Person, but in the materially embodied state the jīvātmā is not
conscious of this relationship, and it is consequently one-sided. In this con
dition the jīvātmā is not directly aware of the L o r d , and either ignores H i m
or appeals to H i m as a vaguely conceived supplier of material needs.
The fundamental postulate of bhakti-yoga is that this is an abnormal
state of affairs. Since the jīvātmā and Krsna are of the same qualitative na
ture, there is a natural symmetry between their respective personal char
acteristics and tendencies. In the Bhagavad-gītā it is stated that Krsna is
constitutionally the dearmost friend of all other conscious beings, and that
He is always caring for them out of natural concern for their welfare. 2 Simi
larly, the jīvātmā has a natural tendency to care for the happiness and well-
being of Krsna, and in a state of pure consciousness the jīvātmā serves
Krsna without desire for personal profit. In this state a loving reciprocal
The Process of Bhakti-yoga 215
relationship develops between the person and Krsna. One secondary con
sequence of this relationship is that by directly contacting the Supreme
Person, the person is placed in touch with the source of all knowledge.
The goal of bhakti-yoga is to purify the person's consciousness so that his
natural relationship with the Supreme can be reawakened. This comes
about through the performance of practical devotional service to Krsna.
Just as a lame person can regain the ability to walk by practicing walking,
so a person in material consciousness can revive his relationship of loving
service to Krsna by actually beginning to practice such service. This can be
accomplished if an initial link can be established that enables the person to
actually serve Krsna through the activities performed by his material body.
A number of important considerations are involved in establishing this
link, and we will discuss these briefly, one at a time.
First we shall discuss the bearing of a person's inner attitudes on his
chances for success in the search for knowledge. The world view of modern
science rests on the idea thatjiature is a product of impersonal processes
that lie within the reach'bf human understanding. Following this idea, *
many scientists view nature as a passive object of conquest and exploita-
tion, and they try to extract the secrets of nature forcibly by the power of
their minds and senses. The theories of modern science are consonant with ,
a domineering and aggressive attitude, and it can be argued that their
development has been strongly influenced by a desire to accommodate
such an attitude.
In contrast, bhakti-yoga is based on the idea that nature is the product of
a supreme intelligence that is beyond the understanding of the human
mind. The approach of bhakti-yoga is not to dominate this intelligence, but
rather to cooperate with it. It is not possible for a person to acquire
real knowledge about Krsna by the power of his limited mind. The key
to bhakti-yoga is that by the mercy of Krsna, such knowledge is readily
available to a person who approaches H i m with a sincerely favorable
attitude.
The quality of this attitude is indicated in the following statement spoken
by Krsna to Arjuna in the Bhagavad-gītā:
Always think of Me and become My devotee. Worship Me and offer your
homage unto Me. Thus you will come to Me without fail. I promise you this
because you are My very dear friend. 3
If a person maintains an inimical or aggressive attitude toward the absolute

truth and regards it as a field of conquest for his mind, then he will have
to depend completely on his ordinary sensory and mental powers in his
search for knowledge. But if the person adopts a genuinely agreeable and
favorable attitude toward the absolute, then by the mercy of the absolute,
the person's internal and external circumstances will gradually be so ad
justed that absolute knowledge becomes accessible to him. The essential
element is the change in attitude. In the beginning a person may have only
the vaguest conception of what the absolute truth may be, but if he adopts
a truly positive attitude toward the absolute, then he will eventually be
able to reciprocate personally with the absolute in a mutual relationship of
love and trust.
This brings us to our second consideration. If a person is initially limited
to his ordinary bodily senses as sources of information, then how can he
make the first step in obtaining transcendental knowledge? Also, if his
ultimate objective is to serve the transcendental Supreme Person, then
how can he do this when his activities are limited to the manipulation of
matter? The answer to these questions is that Krsna is able to reciprocate
with an embodied jīvātmā in two important ways—internally as the all-
pervading Supersoul, and externally through the agency of another em
bodied person who is already connected with H i m in a transcendental
relationship.
Such a person is known as a guru, or spiritual master. The guru is
described in the following way in the Bhagavad-gītā:
Just try to learn the truth by approaching a spiritual master. Inquire from him
submissively and render service unto him. The self-realized soul can impart
knowledge unto you because he has seen the truth. 4
Since the guru is in direct contact with Krsna, he can act as Krsna's repre
sentative. The guru can make information about Krsna available to the
people in general through the medium of the written and spoken word, and
he can also accept service on behalf of Krsna. According to the system of
bhakti-yoga, a person can begin to serve Krsna by accepting a genuine
guru, hearing about Krsna from him, and rendering service to him. Service
to the guru is accepted by Krsna as direct service to Himself, and He recip
rocates by enlightening the person with the knowledge that he needs to
make further advancement on the path of bhakti-yoga.
The process of bhakti-yoga is summed up by the following statement in
the Caitanya-caritāmrta:
Krsna is situated in everyone's heart as caitya-guru, the spiritual master
within. When He is kind to some fortunate conditioned soul, He personally
gives that person lessons in how to progress in devotional service, instructing
the person as the Supersoul within and the spiritual master without. 5
In the initial stages of this process, the aspiring candidate depends almost
Faith, Subjectivity, and Verifiability 217
entirely on the guidance supplied to him externally through the guru. How-
ever, through service to the guru, the candidate's link with Krsna is estab-
lished, and his own natural relationship with Krsna is gradually awakened.
9.2 Faith, Subjectivity, and Verifiability
At this point we should make a few observations about the role of faith in
bhakti-yoga. It is often said that religion is based either on subjective ex-
periences that cannot be verified by other persons, or on received doctrines
that cannot be verified at all. The charge is therefore made that religion is a
matter of blind faith. We should stress that this charge does not apply to the
process of bhakti-yoga, for bhakti-yoga is based on verifiable observation.
It is true that the realizations attained by this process cannot be verified by
means of ordinary sense perception. But they can be verified by other
persons who are also capable of exercising their higher sensory capacities.
In Chapter 3 we illustrated this point by the example of two seeing per-
sons observing a sunset in the presence of a congenitally blind person. The
seeing persons are able to discuss what they see, and each will feel confi-
dent that both he and the other person really are witnessing a sunset. If
necessary, they can confirm this by consulting other seeing persons. In
contrast, the congenitally blind person cannot verify the existence of the
sunset, and he is probably unable to form a realistic conception of what it
would be like to see it. He can either accept the existence of sunsets on
blind faith, reject their existence with equal blindness, or else declare '
himself to be an agnostic.
One might say that it is unfair for a few people to lay claim to knowledge
that can only be obtained by methods not available to people in general.
However, this charge is actually more applicable to certain fields of
modern science than it is to bhakti-yoga. For example, physicists use multi-
million-dollar particle accelerators and elaborate techniques of mathe-
matical analysis to demonstrate the existence of certain "fundamental"
particles. Yet the common man has no access to such expensive equipment,
and he does not have the knowledge needed to properly use it. Since these
assets are very difficult to acquire, the common man has no choice but to
accept the findings of the physicists on faith. Nonetheless, the physicists are
confident that they can verify one another's observations, and they would
not accept the charge that their conclusions are invalid because they cannot
be checked by laymen.
For a given class of observations to be considered objective, the general
rule is that there must be a group of responsible people who are capable of
verifying them. These people must agree on a clear theoretical understand

ing of what observations are to be expected and how they are to be inter
preted. Modern physics is based on such a group of experts, and the same
can be said of the process of bhakti-yoga. The system of bhakti-yoga is
maintained and propagated by a disciplic succession of teachers, or gurus,
who have reached a high platform of personal realization. These teachers
adhere to a standard body of knowledge that is contained in books such as
the Bhagavad-gītā, and their conclusions and conduct can be checked by
the larger community of realized persons, or sadhus. The higher realiza
tions of the process of bhakti-yoga can be discussed and evaluated by the
qualified sadhus just as readily as the findings of experimental physics can
6
be evaluated by expert physicists.
Since bhakti-yoga is based on verifiable observation, it is not dependent
either on blind faith or on speculative argumentation. Yet faith is required
in any difficult undertaking, and the process of bhakti-yoga is no exception.
For example, before studying modern chemistry the prospective student
must have faith that the many experiments on which the subject is based ac
tually work. He cannot know in advance that they will work, but without
faith in this he will not be motivated to carry out the arduous work needed
to master the subject. Normally, the student will begin with a certain
amount of initial faith, and this faith gradually will grow as he acquires
more and more practical experience. The same gradual development of
faith occurs in bhakti-yoga.
Perhaps the main reason for the widespread casting aside of religion
as "blind faith" is that many systems of theistic thought are not backed up
by any verifiable, direct interaction with the Supreme Person. We may
ask why this should be if the Supreme Person is as readily accessible as we
have indicated in our discussion thus far. The following statement in the
Bhāgavata Purāna gives an interesting answer to this question:
The great thinkers can know Him [Krsna] when completely freed from all
material hankerings and when sheltered under undisturbed conditions of the
senses. Otherwise, by untenable arguments, all is distorted, and the Lord
disappears from our sight. 7
One of the most important principles of bhakti-yoga is that higher reali-

zation is impossible as long as the material senses are not brought under
control. In the materially conditioned state of consciousness, the jīvātmā is
motivated by the desire to enjoy his material situation, and he is completely
preoccupied with the barrage of stimuli presented by his material senses. In
this state the jīvātmā's sensory channels are overloaded, and he is unable to
perceive the presence of the Supersoul, even though he is constitutionally
The Brain, the Mind, and the Conscious Self 219
capable of doing this. Since the jīvātmā with uncontrolled senses has no
direct access to the Supreme Person, he is prone to indulge in fanciful
speculations that simply lead him further and further from the truth.
9.3 The Brain, the Mind, and the Conscious Self

To understand some of the practical problems involved in controlling
the senses, we must introduce the concept of the material mind. We have
already pointed out that the jīvātmā is a complete, conscious individual
and that, as such, the jīvātmā is inherently capable of carrying out the men
tal functions of thinking, feeling, and willing. Yet the machinery of the
body includes a psychic subsystem that duplicates some of these functions.
This subsystem acts as an intermediate link between the natural senses of
the jīvātmā and the sensory apparatus of the body. Before reaching the
jīvātmā, data from the bodily senses pass through this subsystem and are
a u g m e n t e d and modified by the addition of information representing
various thoughts, feelings, and desires.
This intermediate link consists of two components, one of which is the
brain. In modern science the brain is generally held to be the seat of all
mental functions. According to the BhagavaJ-gītā, however, the mind has
an additional component which in Sanskrit is called the manah, or material
mind, and which is distinct from both the brain and the conscious self. This
material mind serves as a connecting link between the brain and the self. It
is composed of a kind of material energy, and, in principle, it could be
studied by ordinary ernpirical methods. At present there is no widely
accepted scientific theory of the material mind, but some of the findings of
the parapsychologists may provide the basis for such a theory.
It would take us too far afield to discuss the higher physics of the material
mind here, but we will make a few remarks about the functional rela
tionship between the material mind and the brain. According to the
Bhagavad-gītā, the material mind interacts directly with the brain and
body, and the conscious self interacts with the material mind through the
agency of the Supersoul. The relationship between the brain and the ma
terial mind can be compared with the relationship between a computer and„
a computer programmer. Consider a businessman who has programmed
a computer to process his accounts. The computer can be regarded as an
extension of the man's mind that has its own memory and data-processing
facilities. Even though the man is a complete person in and of himself, he
may come to depend heavily on the computer, and thus his ability to con
duct his business affairs may be greatly impaired if the computer is dam-'
aged. Similarly, the brain is a computerlike extension of the material mind,
and even though this material system can function independently of the
brain, it tends to become dependent on the brain for the execution of
certain data-processing operations.
Taken together, the material body and mind can be regarded as a kind
of false self in which the real self rides as a passenger. This false self is not
conscious in its own right. Both the brain and the material mind can be re-
' garded as mechanisms for symbol manipulation comparable with man-
made computers. The "thoughts" of the material mind are mere patterns of
' symbols, and they become represented by actual thoughts only when they
are perceived by the jīvātmā. The tendency of the embodied jīvātmā, how
ever, is to accept the thoughts, feelings, and desires of the material mind
as his own, and thereby to falsely identify himself as the dramatic persona
represented by these patterns of symbols.
Since the material mind is the director of the material senses, control of
these senses can be achieved by controlling the mind. Since most of us have
never made a real effort to practice such control, we may tend to under
estimate both its importance and the difficulties involved in achieving it.
We may obtain some idea of these difficulties if we consider the powerful
role that habits of thought and action play in our normal activities. The
material mind is a reservoir of elaborate programs that govern everything
from gross movements to subtle attitudes, and thus our mental life consists
of a succession of conditioned thoughts and feelings that unfold inexorably
according to their own internal logic and the stimuli of the senses.
Since we normally tend to identify the material mind with the self, we
have no real idea of what it would be like to be free from this endless torrent
of mundane images and associations. This is indicated, however, by the
following statement in the Bhagavad-gītā:
For one who has conquered the mind, the Supersoul is already reached, for he
has attained tranquility. To such a man happiness and distress, heat and cold,
honor and dishonor are all the same. 8
Once the material mind is under control, the natural senses of the jīvātmā
are free to directly perceive the Supreme Person.
9.4 The Positive and Negative Injunctions of Bhakti
In bhakti-yoga, control of the material mind and senses is achieved by

the following of certain positive and negative injunctions. The negative
injunctions involve the restriction of activities that tend to agitate the ma
terial mind and distract the jīvātmā from the process of self-realization.
The Positive and Negative Injunctions ofBhakti 221
The most fundamental of these injunctions prohibit indulgence in intoxica

tion, meat-eating, illicit sexual affairs, and exploitative business enter-,
prises. We do not have sufficient space here to discuss the psychological
dynamics of these activities in detail, but we note that they tend to cause
persons who engage in them to become more and more preoccupied with
the actions and reactions of their material senses.
In many scientific experiments the physical conditions in the experimen
tal apparatus have to be carefully adjusted if the experiment is to be a suc
cess. The process of bhakti-yoga is an experiment in which the body and
material mind are the experimental apparatus, and in which the negative
injunctions are neccessary (but not sufficient) conditions for the experi
ment's success. These injunctions are essential. If a person neglects them
he will not be able to free himself from material entanglement, and his
"transcendental" realizations will indeed be nothing more than products
9
of self-deception.
The positive injunctions of bhakti-yoga prescribe activities that directly
engage the individual in service to the Supreme Person, Krsna. The effect
of these activities is to awaken the jīvātmā's natural love for Krsna. As a
corollary to this reawakening, the jīvātmā automatically loses his attraction
to the manifestations of his material mind, for these false theatrical displays
are inherently less interesting than the absolute reality of Krsna. Thus, by
engaging in active service to Krsna, the individual is able to attain the goal
of mental control, and free his senses for further service to Krsna.
The ultimate goal of bhakti-yoga is to serve Krsna directly, and this can
be done if the self is freed from entanglement in the affairs of the material
mind and senses. This freedom, in t u r n , can be readily obtained as a
consequence of performing service to Krsna. This may seem like a vicious
circle, but in practice it corresponds to a gradual process of reciprocal
development. First, the material mind must be brought under moderate
control by adherence to the negative injunctions of bhakti-yoga. Then
the individual engages in practical service to Krsna under the guidance of
the guru, and by Krsna's mercy he attains some realization of the Lord.
As a result, his attachment to the material mind is lessened, and he is able
to engage in further service to Krsna on a higher platform of realization.
This leads the individual to further freedom from material desire, and
further realization of his constitutional nature as a servant of Krsna. The
Bhāgavata Purāna sums up this process and its results as follows:
As soon as irrevocable loving service is established in the heart, the effects of

nature's modes of passion and ignorance, such as lust, desire, and hankering,
disappear from the heart. Then the devotee is established in goodness and he
becomes completely happy. Thus established in the mode of unalloyed good

ness, the man whose mind has been enlivened by contact with devotional ser
vice to the Lord gains positive scientific knowledge of the Personality of
10
Godhead in the stage of liberation from all material association.
9.5 The Process of Sravanam
Service to Krsna can take many forms, but of these the most fundamen
tal are the processes of sravanam and kīrtanam, or hearing and chanting.
To awaken one's relationship with Krsna, it is necessary first to hear about
Krsna, and this is also a form of devotional service. One of the principle
themes of this book has been that the absolute truth is not void, but is full
of variegated attributes. The Supreme Person, Krsna, possesses unlimited
personal qualities, and He also engages in unlimited transcendental ac
tivities in reciprocation with innumerable jīvātmās who live in His associa
tion in a state of pure consciousness. By hearing about the attributes and
pastimes of Krsna, the materially conditioned jīvātmā is reminded of his
own natural relationship with Krsna. This provokes in him a desire to know
more about Krsna, and it simultaneously decreases his attachment to the
affairs of the material bpdy and mind.
The philosophy of bhakti-yoga holds that knowledge of the absolute
must descend directly from the absolute. Krsna is the original source of all
material forms, and He is also the origin of the written information which
forms the external subject matter of bhakti-yoga. This subject matter exists
in the form of scriptures that either are produced directly by Krsna Him
self, or are written by persons who are directly linked with Krsna in a tran
scendental relationship. The Bhagavad-gītā is an example of a scripture
of the former type, and the Bhāgavata Purāna and Caitanya-caritāmrta
are scriptures of the latter type. As we have pointed out, the subject matter
of bhakti-yoga is preserved and disseminated by the community of quali
fied gurus and sadhus, whose role in the regulation of knowledge can be
compared with that of the community of experts in a scientific field.
Unlimited amounts of information about Krsna can be encoded in the
form of sequences of symbols. However, since Krsna is absolute, informa
tion about H i m is different from ordinary information describing configu
rations of matter. In our ordinary experience, patterns of symbols can be
arranged according to the conventions of a language, so as to represent
certain events in a limited region of time and space. When we perceive
this information by hearing or reading, we are able to interpret the coded
patterns, and as a result we become aware of a mental image of the events.
The Process of Sravanam 223
This mental image is something quite different from the events them
selves. However, when the jīvātmā interprets information describing the
Supreme Person, the resulting mental images actually bring the jīvātmā
in direct contact with the Supreme Person. Since Krsna is absolute, ma
terial images and sounds representing Krsna are nondifferent from Krsna
Himself, and this can be understood directly by the jīvātmā. Such under
standing cannot, of course, be simply a matter of material symbol manipu
lation. It directly involves the higher sensory and cognitive facilities of the
conscious self.
Since this point is quite important, let us explore it in greater detail. Ac
cording to the philosophy of the Bhagavad-gītā, nothing is different from
Krsna, yet nothing is Krsna except for His own primordial personality.
Krsna is the cause and the essence of all phenomena, and in this sense all
phenomena are identical with Krsna. Yet the phenomena of this world are
merely external displays projected by Krsna's will, and His real nature is
His eternal personality. As we argued in Part I I , the absolute is highly spe
cific, and therefore Krsna can be represented by certain symbolic patterns
and not others. By means of these patterns, Krsna can introduce Himself to
the conditioned jīvātmā, and thus these configurations are nondifferent
from Krsna in a direct personal sense. By perceiving such configurations,
the jīvātmā is reminded of Krsna, and by Krsna's mercy he is allowed to see
Krsna directly by his own higher vision.
This explanation may convey some idea of how the embodied jīvātmā,
who is normally restricted entirely to material modes of sense perception,
can begin to perceive the transcendental Supreme Person. In this initial
stage, the jīvātmā's perception of Krsna may seem to be completely depen
dent on the interactions of matter, but the essence of his experience is not
material. We can begin, to understand this by considering the ideas that
matter itself is a manifestation of Krsna, and that material perception is
simply a limited, impersonal way of seeing Krsna.
In the highest stage of realization the reciprocation between the jīvātmā
and Krsna has nothing to do with the material manifestation. This relation
ship is not dependent on the material body of the jīvātmā in any way, and it
continues after the body has ceased to exist. According to the philosophy of
bhakti-yoga, the material manifestation represents only a minor aspect of
the total reality. There is a higher realm that is inaccessible to material
sense perception, but is nonetheless full of variegated form and activity.
Here we are concerned with the question of how a materially embodied
person can acquire absolute knowledge, and so we will not try to discuss
this higher realm in detail.
9.6 The Process of Kirtanam
The process of śravanam, or hearing, is complemented by the process of

kirtanam, or glorifying the Lord by singing about His qualities and pas
times, and discussing these with others. We have argued that the process
of bhakti-yoga is scientific in the sense that it is a practical method of ob
taining verifiable knowledge about the absolute truth. In the science of
bhakti-yoga, however, the researcher approaches the absolute with an at
titude of reverence and devotion, in stark contrast to the aggressive and ex
ploitative approach that prevails in modern science. By glorifying Krsna,
the individual can awaken his natural love for Krsna, and once this is done,
Krsna will be fully accessible to him on a personal level.
One important form of kirtanam is the chanting of the names of Krsna.
In Chapter 1 we mentioned that bhakti-yoga involves the chanting of the
mantra,
Hare Krsna, Hare Krsna, Krsna Krsna, Hare Hare

11
Hare Rāma, Hare Rāma, Rāma Rāma, Hare Hare.
The Sanskrit term mantra refers to a pattern of sound that has a purifying
effect on the mind. This particular mantra is composed of three'names of
the Supreme Person. Grammatically, the mantra is in the vocative case,
and its meaning is to address the Lord by calling out His names.
These names are examples of symbolic patterns that directly represent
the Absolute Person, and therefore have an absolute, inherent meaning.
According to the philosophy of bhakti-yoga, Krsna's names are nondiffer-
ent from Krsna Himself, and by chanting and hearing these names, one is
brought into personal contact with Krsna. This can be directly appreciated
by persons whose higher sensory capacities have been awakened. For
others, the chanting of these names purifies their consciousness and
thereby brings about this awakening.
The results of chanting the names of the L o r d can be obtained by the use
of any names that are actually connected with the Supreme Person, and are
not mere concoctions of the material imagination. The significance of this
chanting was described as follows by Śrī Caitanya Mahāprabhu, the great
teacher of bhakti-yoga who appeared in India in the fifteenth century:
O my Lord, Your holy name alone can render all benediction to living beings,
and thus You have hundreds and millions of names, like Krsna and Govinda.
In these transcendental names You have invested all Your transcendental
energies. There are not even hard and fast rules for chanting these names. O
my Lord, out of kindness You enable us to easily approach You by chanting
Your holy names, but I am so unfortunate that I have no attraction for them. 1 2
The Process of Kirtanam 225
The purport of this statement is that due to the blindness caused by his
preoccupation with the material mjnd and senses, the conditioned jīvātmā
will initially feel very little attraction to chanting the names of the L o r d . Yet
by regularly chanting these names and following the regulative injunctions
of bhakti-yoga, this blindness can gradually be cured, and one can attain
the stage of loving reciprocation with Krsna.
Since the ultimate goal of this chanting is the development of love, it
must be carried out with an inner attitude that is compatible with this
emotion. Śrī Caitanya Mahāprabhu goes on to describe this attitude as
follows:
One should chant the holy name of the Lord in a humble state of mind, think
ing oneself lower than the straw in the street; one should be more tolerant
than a tree, devoid of all sense of false prestige, and ready to offer all respect
to others. In such a state of mind one can chant the holy name of the Lord
13
constantly.
It may not be possible for a person who has no direct knowledge of the
Supreme Person to immediately understand what it might mean to love the
Supreme. However, it is possible for him to lay the groundwork for such
understanding by adopting a nonexploitative attitude toward the Supreme
Person and His creation. Indeed, this is the key to the process of bhakti-
yoga. To one who wishes to exploit the Supreme, the Supreme will remain
unknowable. But if one truly gives up the desire for such exploitation, then
the Supreme Person will reveal Himself by His own mercy.
In the letter quoted at the beginning of this chapter, Einstein declared
that his goal was to capture the absolute truth. The absolute truth cannot be
captured forcefully by a minute part of the absolute, but according to the
philosophy of bhakti-yoga, the absolute can be captured by love. Once this
love is attained, direct knowledge of the absolute will be readily available,
but paradoxically, the development of this love is not compatible with the
desire for knowledge or power. Knowledge is indeed a byproduct of the
process of bhakti-yoga, but it cannot be the goal of that process, for the key
to the process itself lies in the fundamental reassessment of one's innermost
goals.
Although this reassessment may be superficially simple to describe, to
carry it out requires a deep insight into one's own psychology. By bringing
the inner self into personal contact with the absolute, the process of bhakti-
yoga enables one to attain this insight. By this means the absolute can be
captured—once all desire to conquer the absolute has been forsaken.
Notes
1. F r e n c h , ed., Einstein, p p . 275-276.

text 5.29, p p . 303-304.
3. Ibid., text 18.66, p. 834.
4. Ibid., text 4.34, p. 259.
5. Kavirāja, Śrī Caitanya-caritāmrta, Madhya-lila, Vol. 8, text 22.48,

p . 351.
6. T h e regulation of the process of bhakti-yoga in the society of realized

sadhus is vividly described in the Śrī Caitanya-caritāmrta. This should
be c o m p a r e d with the regulation of scientific research in the scientific
community.
7. A . C . B h a k t i v e d a n t a Swami Prabhupāda, Śrīmad-Bhāgavatam,

C a n t o 2, Vol. 1, text 2.6.41, p. 347.

text 6.7, p. 315.
9. We stress this point because t h e r e are many watered-down systems of

yoga or meditation in which even the most basic rules for sense control
are neglected. Seeking self-realization through such systems is com
parable to trying to ignite firewood while simultaneously pouring
w a t e r on it.
10. A . C . Bhaktivedanta Swami Prabhupāda, Śrimad-Bhāgavatam,

C a n t o 1, Vol. 1, texts 1.2.19-20, p p . 116-118.
11. T h e chanting of this mantra is r e c o m m e n d e d in the Kalisantarana

Upanisad. (See A . C . B h a k t i v e d a n t a Swami P r a b h u p ā d a , Teachings of
Lord Caitanya, p p . 203-204.)
12. A . C . Bhaktivedanta Swami Prabhupāda, Śrimad-Bhāgavatam,

C a n t o 1, Vol. l , p p . 41-42.
13. Ibid., p. 4 1 .
Appendix 1
A Discussion of Information Theory
The measure of information content used in Chapter 5 is the subject of
a field of study called algorithmic information theory. In this appendix
we will discuss the definitions and theorems from this field that we used in
Sections 5.2 and 5.3. We will present a brief summary of important results.
Additional details can be found in Thompson, (1980), which contains a
study of observation functions from the viewpoint of information theory.
A general review of algorithmic information theory can be found in
Chaitin, (1977).
To define information content we must first establish a fixed computer,
C. We assume that C accepts as programs bit strings P written in a particu
lar programming language similar to BASIC. Let R denote the rational
numbers. A program P is said to compute the n-ary function X: R" > R
on the computer C if the computer calculates X(Y1, . . . , Y n ) as output,
given values Y1, . . . ,Yn ε R as input. (If n = 0 then X is a constant and
there is no input.)
Any suitably powerful programming language can be used to define a
measure of information content. However, we have made certain assump
tions about the programming language that simplify the algebra of algo
rithmic information theory. We will introduce some of these assumptions
as we go along, and the remainder can be found in (Thompson, 1980). We
will assume that programs are written in an alphabet of 64 characters, each of
which can be encoded as a 6-bit string. These include the symbols 0 9
and A, . . . ,Z. We assume that the language has a system for encoding any
integer 0 x<2n with a string of no more than n + 6 log 1 0 (n) + 12 bits. We
also assume that the computer can handle rational numbers and integers
with arbitrarily many significant digits.
Let X: R" > R be an n-ary function that is computable by C. The in
formation content of such a function has been defined by A . I . Kolmogorof
(1968), G. Chaitin (1977), and others to be
L(X \C) = min{l(P) : P computes X on C} (25)
Here l(P), the length of P, is the number of bits in the bit string P.
We can see from this definition that information content is not an abso
lute, but is dependent on the particular computer, C. Yet for any C the
227
228 Appendix 1
vast majority of X's have a large information content. There can be no

more than 2 programs P of length less than or equal to k. It follows that
k
the number of integers 0 x<2 with L(x|C) k is no more than 2 ,

n k
n
even though there are 2 numbers in this range. The conventions of our
programming language for the coding of integers imply that
L(x |C) 36 + 6 log (n) + n 10 (26)

and we can thus conclude that most binary integers of n bits have an
information content of approximately n.
The computer C can be modified by the addition of m externally supplied
functions or constants F . . . ,F that can be called on by name in pro-
1, m
grams. This modified computer can be called C(F . . . ,F ), and we 1, m
write L(X |F . . . ,F ) = L(X \C(F . . . ,F )) and also L(X) = L(X |C).

1, m 1, m
L(X |F . . . ,F ) is a measure of the amount of additional information

1, m
needed to specify X, given that F . . . ,F are known. The programming

1, m
language is designed so that
The following basic theorem expresses the relationship between infor-

mation content and the probabilities of events in a mathematical model
of a physical system.
P r o p o s i t i o n 1.
Suppose that M is a computable function mapping the non-negative
integers into the non-negative rational numbers. Suppose that M satisfies
where c= 163 + 5.42r.
This theorem is proven in Thompson, (1980), and it also holds if we

replace L(X |M) in (30) by L(X | M , F . . . ,F ). 1 m
(Corollary 1.
Suppose that M is the same as in Proposition 1. Then, given the same
A Discussion of Information Theory 229
assumptions on X,
C + V + L ( M ) - L ( X )
M(X) 2 (31)
This corollary follows from Proposition 1 and (28) with m = 0 and F =M. l
Inequality (18) of Section 5.3 follows from (31) with r = 3 0 . The require-
30 9
ment that r = 3 0 simply means that (18) applies only for L(X)<2 10 .
Let X , . . . ,X be non-negative integers, and let G be a function
1 m
mapping non-negative integers to non-negative integers. Suppose that

G(X )= . . . =G(X ) — Y. In Section 5.3 we explained how G could be
l m
interpreted as a process of observation, and how Y could be interpreted as

a description of common observed features shared by all the X 's. There we j
also pointed out that L(Y |G) could be taken as a measure of the amount
of observable information that is derived by G from the X 's, but is not j
simply an artifact of G. We shall define the amount of information shared

by X , . . . ,X to be
1 m
(32)
Here X stands for the set, {X ,. . . , X } . 1 m
J ( X ) represents the maximum amount of information that can be ob-

served in each of the X 's by some process of observation. We can simi-
j
larly define J(X |F . . . ,F ) by using L(Y | G , F . . . ,F ) in place of

l n l n
L(Y |G) in (32). This represents the amount of shared information in the
X 's that is independent of F . . . ,F .
j l n
Proposition 2.
Let M be a probability measure on the integers, 0 x < N . Suppose that
M takes on rational values, and that N can be expressed in 15 or fewer
characters.
If X is a set of integers, 0 X . . . ,X <N, and J(X|M)<2 for some
1 m
r
integer r > 0 , then
(33)
where c'=337 +3.62r.
This is proven in T h o m p s o n , (1980), and it also holds w i t h
J(X |M,F . . . ,F ) in place of J(X |M).
1 n
Corollary 2.
With the same assumptions as in Proposition 1,
(34)
230 Appendix 1
This corollary also follows using (28). Inequalities (24) and (22) in Sec-
tion 5.3 follow directly from (33) and (34). (Again, we choose r=30.) In
Thompson, (1980) the properties of J ( X ) are analyzed, and a number of
examples of observation functions are discussed.
In Section 5.2 we argued that if a long bit string has a low information
content, then there must exist a strong interdependence among its parts.
Now we shall show how this can be formally established. Let X be a m
bit string of length hm, and suppose that X is divided into substrings
m
Y , . . . , Y of length h. If we treat the Y 's as binary numbers we can write

1 m k
X =1 and
0
h
X k + 1 = 2 X +Y
k k + 1 (35)
for each 0 k<m. (Since X may begin with leading zeros, we place a 1 in
m
front of it to convert it into a binary number.)

Let N be a very large fixed number. We say that the bit string X is an
initial segment of the bit string X' if X' consists of X followed by n 0
additional bits. We define I(X) to be the number of bit strings X' such
that L(X') N and X is an initial segment of X'.
Now let w be a bit string (which may have leading zeros). We define
F(h,w,X) to be the wth integer Y in numerical order for which 0 Y<2 h
and
h -1(w)
I(2 X+ Y)>2 I(X) (36)
This F can be called the generating function. (F was used in Section 5.2
with the parameter h suppressed for convenience.)
Our application of F in equation (11) of Section 5.2 was based on the
following proposition.
Proposition 3.
Let Y . . . ,Y be as in (35), and suppose that L(X )
1 m m N. Then there
are bit strings w , . . . ,w such that
1 m
A Discussion of Information Theory 231
According to (39), inequality (42) is satisfied by Y= Y . Therefore Y is the

n n
wth number from 0 to 2 -1 satisfying (42), where w lies in the interval,

h
1 w 2 —1. We can produce the desired bit string w by adding leading

Kn
n
zeros to this w, if necessary, so that its length equals K . Then (38) will
n
be satisfied.
To show (37), note that I(X ) is no greater than the number of strings X'
0
with L(X') N, and that this is no greater than 2 . Now let n = L(X ) N.
N
m
There must be a program P of length n that computes X . The conventions

m
of our programming language can be arranged so as to allow us to construct

a program Q of the form Q=PA, where A is an arbitrary string of N—n bits.
The idea here is that when executing Q, the computer will refer only to the
instructions in P, and will never refer to A. Thus A can be chosen in any of
2 different ways, and it follows that,
N-n
We obtain (37) by combining (43), (40), and I(X ) 2 . Q . E . D .

0
N
We have seen that the generating function F can be defined very simply.
However, an inspection of the definition will show that it is not possible to
compute F in practice. In fact, as matters stand, F cannot be computed at
all. It can be shown that even though L(X) is a precisely defined function,
there does not exist an algorithm that will compute L(X) for any value of
X. (This is discussed in (Chaitin, 1977).) This means that L(X) is what is
known as an uncomputable or "nonrecursive" function. Since F is defined
in terms of L, it is not surprising that F is also uncomputable.
232 Appendix 1
It is interesting that we can unambiguously define numbers and functions

that are theoretically impossible to compute. We will not have to consider
the implications of this here, however, for we can easily modify our defini-
tions in such a way that F becomes computable. One way of doing this is to
place a very large cutoff on the running time of programs on our computer,
C. If we stipulate that any program running over this limit immediately
prints a 0 and stops, then both L and F become formally computable, and
all our considerations in Sections 5 . 1 , 5 . 2 , and 5 . 3 remain essentially unaf-
fected. (We must set the cutoff high enough so that the program for M(X)
can complete its calculations.)
Nonetheless, even though F is rendered computable by this device, the
number of steps needed to calculate F is so large that we cannot hope to
carry them out in practice. One might therefore ask, "What is the value of
considering F if this function cannot be computed?" In brief, our answer is
as follows: If we choose an appropriate cutoff on the running time of pro-
grams on C, we can write a very simple program that will calculate F (but
not within the time limit imposed by this cutoff). This program involves a
few simple arithmetical operations that are repeated over and over again
according to a simple pattern. Now, suppose that a detailed description of
biological form has a low information content. Then we can apply Proposi-
tion 3 and show that successive sections of this description can be generated
by applying F to successive code strings w , . . . ,w . AS we described in
1 n
Section 5 . 2 , very elaborate segments of the biological description can be

generated in this way by the addition of miniscule amounts of information
in the form of the strings w . j
Our question is, " H o w can the function F reel off successive descriptions
of complex organs and biological processes under the guidance of such mi-
nute amounts of information?" If we suppose that F can do this, then we
must conclude that the mere repetition of a few simple computational
steps, if prolonged over a long enough period, will generate the kind of
complex patterns found in living organisms. Furthermore, given the direc-
tion provided by the w -'s, the specific patterns of life as we know it will be
j
produced, and no others. This would seem to attribute very remarkable

powers to the elementary operations of arithmetic. Although we cannot
rigorously prove that F does not have these remarkable properties, we
think it worth our while to seriously consider the consequences that follow
if, in fact, it does not have them.
Appendix 2
Information Content of the Laws of Chemistry
In this appendix we shall discuss the information content of the physical
laws thought to govern both chemical reactions and the interactions be-
tween matter and electromagnetic radiation. We will be particularly con-
cerned with the laws of nonrelativistic quantum mechanics summarized in
Figure 1 of Chapter 5. These laws apply to a physical system containing M
particles with position coordinates Q , . . . , Q . The particles consist of
1 M
electrons plus several different species of atomic nuclei. We can assume

that the particles are situated within a very large box containing a radiation
field characterized by the parameters q (n = 1,2, 3, . . . ) .
n
In Figure 1, equation (a) is the Schr dinger equation, and ¥ is a wave

function defining the state of the system. In equation (b), the Hamiltonian
H is expressed as a sum of terms representing different kinds of physical in-
teractions. Reading from left to right (and from top to bottom), the first
term represents the propagation of electromagnetic radiation in space.
The second term represents the kinetic energy of the particles, and the next
two terms describe the interaction between the charged particles and the
electromagnetic field. The fifth term represents the interaction between
the electromagnetic field and the spin, or magnetic moment, of the parti-
cles. Finally, the sixth term describes the electrostatic and gravitational
interactions of the particles.
To get some idea of the information content of the laws of physics, we
have written a program for the numerical solution of the Schr dinger equa-
tion using the Hamiltonian of Figure 1. It is not possible, of course, to run
such a program on an actual computer, for the required computations
would tax the powers of the fastest computers for even a very small number
of particles. Nonetheless, it is interesting to see how the laws of physics ap-
pear when their abstract definitions are expressed in terms of elementary
arithmetical operations.
In Appendix 1 we defined information content in terms of a fixed pro-
gramming language. This language is written in a 64-character alphabet
that includes 0 9 and A, . . . ,Z. Special symbols are also included
for
IF THEN GOTO FOR TO NEXT PRINT
233
234 Appendix 2
These symbols are adopted from the B A S I C programming language, and

their meanings and syntactical rules are essentially the same as in that lan-
guage. There are also a few other symbols for special purposes, including a
symbol for the operation, I N T , of rounding a number down to an integer.
These symbols designate the primitive operations of the language, and
all other operations, such as the extraction of roots, must be defined in
terms of these elementary symbols. (The language is described in detail in
Thompson, (1980).)
In this language, a program for the numerical solution of the equations in
Figure 1 can be written in about 1,738 characters, or less than two-thirds of
a page. In addition to programming instructions, a table of data is also
needed to define various physical constants, such as the charges and masses
of the different kinds of atomic nuclei. If we allow 1,062 characters for
this data, the total programming for the representation and solution of the
quantum mechanical equations comes to one page of 2,800 characters.
In writing this program we have made a number of assumptions. First of
all, we assumed that the box containing the physical system has what are
known as "periodic" boundary conditions. This makes it possible to ex-
pand the electromagnetic vector potential A in a standard way as a sum of
sine and cosine terms, A . We assumed that the box is very large, and that
n
the particles are bound by gravitation near the box's center. (This can be
accomplished by introducing a fixed gravitational potential.) These as-
sumptions mean that the boundary of the system can be regarded, in effect,
as an infinite vacuum. Since our model does not allow for nuclear inter-
actions, a source of heat and light is necessary, and this can be provided by
introducing a fixed source of electromagnetic radiation.
We assumed that electrons have a spin of V2, but we neglected the spins of
the nuclei, which were treated as charged point masses. We did not intro-
duce any relativistic correction terms, and we used a cutoff to bound the
Coulomb potential, 1/r, for r<e. We also introduced a frequency cutoff by
using a finite number of modes of oscillation, q . Finally, we introduced a
n
finite grid for the approximation of derivatives, and we placed bounds on

the number of significant figures to be carried in various calculations.
If we contemplate these assumptions, it becomes clear that our model
cannot be taken as a final statement about the laws underlying chemistry.
Unfortunately, when we examine the matter closely, we find that it is very
difficult to adequately formulate such laws. It is very hard to say whether or
not any particular change in our assumptions would significantly affect the
validity of our model as a representation of chemical and biological proces-
ses. For example, how many significant figures should we include for the
charge and mass of an electron? Should we include relativistic effects?
Information Content of the Laws of Chemistry 235
Should we go further and include other physical interactions, such as the

weak force? ( A t least one author has speculated that the weak force may
1
have biologically significant effects. )
At present, no mathematically consistent theories have been formulated
that take the weak and nuclear forces into account, and there are also se-
rious mathematical difficulties plaguing the relativistic quantum theory
2
of electromagnetism. Even the nonrelativistic quantum theory that
3
we have considered here is beset with controversial problems. Some of
these problems are discussed in Chapter 3. Another was raised by the
physicist Eugene Wigner, who argued that a self-reproducing organism is
4
not possible in a quantum mechanical system.
These theoretical problems are accompanied by the practical problem
that it is very difficult to analyze complex chemical reactions using the quan-
5
tum theory. As a result, many researchers accept quantum mechanics on
faith, but use reaction-diffusion equations to model biochemical systems.
Yet as we pointed out in Section 5.3, such models have fundamental draw-
backs that make them incapable of adequately representing organisms.
Here we have decided to use the quantum mechanical equations of Chap-
ter 5, Figure 1 simply to provide a general idea of the amount of information
entailed by the known laws of physics and chemistry.
If we examine the theory of evolution, we find that three basic principles
are generally invoked. These are (1) that self-reproducing organisms of
great complexity are physically possible, (2) that these organisms can
change by random mutation, and (3) that the organisms will be cullecfby
natural selection. One might therefore argue that any laws sufficient to
guarantee (1), (2), and (3) could serve as the basis for a model of evolution.
As Wigner's arguments indicate, it is not clear at present whether or not
these three principles are guaranteed by the known laws of quantum
mechanics. However, as we pointed out in Section 5 . 1 , John von Neumann
has devised a kind of mathematical model that does provide for (1), (2),
and (3). His cellular automata are nonphysical, for they involve many
finite-state automata situated in the cells of a two-dimensional lattice.
Yet they can represent self-reproducing "organisms" (universal Turing
machines) of any degree of complexity.
We have therefore written a program for the M(X) function of a particu-
6
larly simple cellular automaton model due to E.F. C o d d . This model can
also represent self-reproducing universal Turing machines, and it can easily
be modified to provide for random mutations. The model entails an ini-
tially empty lattice that gradually fills up with a "soup" of randomly dis-
tributed states. The program for M(X) requires 1,856 characters, or about
two-thirds of a page.
236 Appendix 2
As we have pointed out in Section 5.3, the probability for the evolution
of recognizable higher forms is practically zero in a model of such low in
formation content. Even if the model were started with an initial popula
tion of primitive self-reproducing automata, no higher forms recognizable
to us could be expected to evolve. A process of natural selection is to be
expected in a population of self-reproducing automata that must compete
for space and resources. One might therefore imagine that some kind
of "higher intelligence" might gradually evolve. Yet if this happened,
the "intelligence" would be so alien that it would be meaningless to call
it by this name. In general, no representative of a class of forms is likely
to evolve if the class is characterized by a description of high information
content.
As a final point, let us consider how M(X) might be defined for our quan
tum mechanical model. To define M(X) we would first have to devise a
coding scheme whereby molecular configurations could be represented by
integers, X. A necessary condition for two atoms to be bonded together in
a molecule is that their nuclei should be situated within a certain distance of
one another. This distance will depend on the types of atoms and the type
of bond being considered. The pattern of bonds in a particular molecule
can thus be represented as a network of internuclear distances.
Let us suppose that the nuclei in our quantum model have coordinates
Q1 . . . ,QN where N<M. Using our coding scheme, we can define a
function B x ( Q 1 . . . ,Q N ) which will equal 1 if some subset of the nu
clear coordinates Q 1 . . . ,Q N satisfy the spacing requirements for the
molecule specified by X. This function will otherwise be set equal to 0.
The function B x can be used to estimate the probability that the mol
ecule described by A'exists within the physical system. If the system is in
the quantum mechanical state, V, let
(44)
where the integral is taken over all the variables of Since Bx defines a
necessary (but possibly insufficient) condition for the existence of the
molecule X, M(X,Ψ) gives an upper bound on the probability that this
molecule exists in the system.
A given arrangement of Q1 . . . ,QN can represent only a limited
number of molecular configurations. There should be a number T for
which
(45)
Notes 237
for all arrangements of Q 1 . . . ,Q N .

To obtain an estimate of T, we shall make a further restriction on B x .
We require that B X (Q 1 . . . ,Q N ) shall equal 1 only if some of these coor
dinates satisfy the conditions for X but are not within bonding distance of
any of the other Qj's. Then Bx= 1 will mean that the molecule represented
by X is present, but is not part of a larger molecule. This means that each
Qy can be part of at most one configuration corresponding to an X with
Bx(Q1 . . . ,QN) =1. Therefore, T N.
The function M(X,Ψ) can be specified using (44) and the definition of Ψ
in terms of natural laws, initial conditions, and boundary conditions. We
can also calculate the probability of finding X in a system described by
a mixture of states. If the mixture consists of the states Ψ j with probabil
ity a J , then this probability is bounded by Σa j M(X,Ψj). Thus, the total
information content of the function M(X) is given by
L(M) L(initial conditions) + L(boundary conditions) (46)

+ L(natural laws) + L(B) + L(t) + constant.
The Bx we have been considering can be expressed in less than a half page
of coding, and thus L(B) need be no more than half a page, or 8,400 bits. In
Section 5.2 we allotted a full page for this term so as to allow for a more
sophisticated function, Bx.
The constant represents the number of symbols needed to express (44),
plus some other odds and ends. The time must be considered in our esti
mate of L(M) since it enters into the calculation of the state of the system.
L(t) will be negligible, however, for times ranging from 0 to billions of
billions of years. We have seen that L (natural laws) is about one page
in our model. If we allow four pages for the definitions of the initial and
boundary conditions, we obtain the upper bound on L(M) given in (16) of
Section 5.3.
Notes
1. G a r e y , "Superweak Interactions and the Biological Time D i r e c t i o n , "

p p . 1-5.
2. Since its inception in the late 1940s, the theory of relativistic q u a n t u m

electrodynamics has been beset by serious mathematical difficulties.
T h e standard m e t h o d of resolving these difficulties has been a proce
d u r e called renormalization. This procedure has enabled physicists
to m a k e certain calculations that have agreed remarkably well with
238 Appendix 2
experimental measurements. However, renormalization has not elimi-

nated the mathematical difficulties plaguing the theory, and the physi-
cist P. A . M . Dirac has suggested that the renormalization theory "will
not survive in the future, and that the remarkable agreement between
its results and experiment should be looked on as a fluke." (Dirac,
"Evolution of the Physicist's Picture of N a t u r e , " p. 50.)
3. Wigner, "Epistemological Perspective on Q u a n t u m T h e o r y . "
4. Wigner, " T h e Probability of the Existence of a Self Reproducing U n i t , "

p p . 231-238.
5. Wigner, "Epistemological Perspective on Q u a n t u m T h e o r y , " p. 374.
6. C o d d , Cellular Automata.
Bibliography
A d l e r , J. "Is M a n a Subtle A c c i d e n t ? " Newsweek, 3 N o v e m b e r 1980,

p p . 95-96.
A n d r e w s , H e n r y N. Jr. Studies in Paleobotany. New York: J o h n Wiley and

Sons, 1961.
Axelrod, Daniel I. " E v o l u t i o n of the Psilophyte Paleoflora." Evolution.

Vol. 13, J u n e 1959, p p . 264-275.
Axelrod, Daniel I. " T h e Evolution of Flowering P l a n t s . " In Evolution

After Darwin. Vol. 1, The Evolution of Life. E d i t e d by S. Tax. Chi
cago: University of Chicago Press, 1960.
Bell, Eric T. Men of Mathematics. N e w York: Simon and Schuster, 1937.
Bell, John S. " O n the Problem of H i d d e n Variables in Q u a n t u m Me

chanics." Reviews of Modern Physics. Vol. 38, n o . 3, July 1966,
p p . 447-452.
Berg, H o w a r d C. " H o w Bacteria Swim." Scientific American, August

1975, p p . 36-44.
B h a k t i v e d a n t a Swami P r a b h u p ā d a , A . C . Bhagavad-gītā As It Is. Sanskrit

text, translation, and c o m m e n t a r y . N e w York: Collier Books, 1972.
Bhaktivedanta Swami P r a b h u p ā d a , - A . C . Śrī īśopanisad. Sanskrit text,

translation, and c o m m e n t a r y . Los Angeles: Bhaktivedanta Book
Trust, 1974.
Bhaktivedanta Swami P r a b h u p ā d a , A . C . Śrīmad-Bhāgavatam of Krsna-

Dvaipāyana Vyāsa. Sanskrit text, t r a n s l a t i o n , a n d c o m m e n t a r y .
C a n t o s 1-10 (30 vols.) Los Angeles: B h a k t i v e d a n t a B o o k Trust,
1972-1980.
Bhaktivedanta Swami P r a b h u p ā d a , A . C . Teachings of Lord Caitanya. Los

Angeles: B h a k t i v e d a n t a Book Trust, 1974.
Bhaktivedanta Swami P r a b h u p ā d a , A . C . The Nectar of Devotion: A

Summary Study of Śrīla Rūpa Gosvāmī's "Bhakti-rasāmrta-sindhu".
Los Angeles: Bhaktivedanta Book Trust, 1970.
239
240 Bibliography
B o h m , D a v i d . Causality and Chance in Modern Physics. London:

Routledge and Kegan Paul, 1957.
B o h m , David and B u b , J. "A Proposed Solution of the M e a s u r e m e n t

P r o b l e m in Q u a n t u m Mechanics by a H i d d e n Variable T h e o r y . "
Reviews of Modern Physics. Vol. 38, n o . 3, July 1966, p p . 4 5 3 - 4 6 8 .
Brush, Stephen G. "Should the History of Science be R a t e d X ? " Science.

Vol. 183, 22 March 1974, p p . 1164-1172.
B u d d e n b r o c k , Wolfgang von. The Senses. A n n Arbor: University of

Michigan Press, 1958.
Capra, Fritjof. The Tao of Physics. New York: Bantam B o o k s , 1975.
Chaitin, Gregory G. "Algorithmic Information T h e o r y . " IBM Journal of

Research and Development. Vol. 2 1 , n o . 4, July 1977, pp. 350-359.
C o d d , E. F. Cellular Automata. New York: Academic Press, 1968.
Cozzarelli, Nicholas R. " D N A Gyrase and the Supercoiling of D N A . "

Science. Vol. 207, 29 February 1980, p p . 953-960.
Crick, Francis. "Split G e n e s and R N A Splicing." Science. Vol. 204,

20 April 1979, p p . 2 6 4 - 2 7 1 .
D a n e r i , A . ; Loinger, A . ; and Prosperi, G . M . "Further R e m a r k s o n

the Relations Between Statistical Mechanics and Q u a n t u m Theory
of M e a s u r e m e n t . " // Nuovo Cimento. Vol. 44B, n o . 1, 1966,
pp.119-128.
D a n e r i , A . ; Loinger, A . ; and Prosperi, G . M . " Q u a n t u m Theory of

M e a s u r e m e n t and Ergodicity Conditions." Nuclear Physics. Vol. 33,
1962, p p . 2 9 7 - 3 1 9 .
Darwin, Charles. On the Origin of Species. New York: A t h e n u m , 1972.
Darwin, Charles. The Life and Letters of Charles Darwin. Vol. 1. New
York: D. A p p l e t o n , 1896.
Dewitt, Bryce S. " Q u a n t u m Gravity: the New Synthesis." In General Rela-

tivity. Edited by S. W. Hawking and W. Israel. Cambridge: Cambridge
University Press, 1979.
Dewitt, Bryce S. " Q u a n t u m Mechanics and Reality." Physics Today,

September 1970, p p . 3 0 - 3 5 .
Bibliography 241
Dickerson, Richard E. "X-ray Analysis and Protein Structure." In The

Proteins, Vol. II. Edited by H. Neurath. New York: Academic Press,
1964.
Dirac, P . A . M . " T h e Evolution of the Physicist's Picture of N a t u r e . "

Scientific American, May 1963, p p . 4 5 - 5 3 .
Dixon, Malcolm and Webb, Edwin C. Enzymes. 3d ed. New York:

Academic Press, 1979.
Dobzhansky, Theodosius. "Darwinian Evolution and the Problem of

Extraterrestrial Life." Perspectives in Biology and Medicine. Vol. 15,
n o . 2, Winter 1972, p p . 157-175.
Dobzhansky, Theodosius. "From Potentiality to Realization in Evolu-

tion." In Mind in Nature. Edited by J . B . C o b b , Jr. and D . R . Griffen.
Washington, D . C . : University Press of America, 1978.
D o d s o n , Edward O. and D o d s o n , Peter. Evolution, Process and Product.

New York: D. Van Nostrand C o . , 1976.
Eddington, A r t h u r . The Philosophy of Physical Science. New York:

Macmillan C o . , 1939.
E d w a r d s , Paul, ed. The Encyclopedia of Philosophy. Vol. 5. New York:

Macmillan Publishing C o . and The Free Press, 1967.
Eigen, Manfred. "Selforganization of Matter and the Evolution of Bio-

logical Macromolecules." Die Naturwissenschaften. Vol. 10, October
1971, p p . 4 6 5 - 5 2 3 .
Eigen, Manfred and Schuster, Peter. "The Hypercycle," Parts A, B, and

C. Die Naturwissenschaften. Vol. 64,1977, p p . 541-565; Vol. 65,1978,
p p . 7 - 4 1 ; Vol. 65, 1978, p p . 341-369.
Elliot, Alfred M. and Ray, Charles, Jr. Biology. 2d ed. New York:
Meridith P u b . C o . , 1965.
F i n e , T e r r e n c e L. Theories of Probability. N e w York and L o n d o n :

Academic Press, 1973.
Fisher, Ronald. The Genetical Theory of Natural Selection. 2d ed. New
York: Dover, 1958.
F o d o r , Jerry A. " T h e Mind-Body P r o b l e m . " Scientific American, January
1981, p p . 1 1 4 - 1 2 3 .
242 Bibliography
F r e n c h , A . P . , ed. Einstein: A Centenary Volume. Cambridge, Mass.:

Harvard University Press, 1979.
Freundlich, Yehudah. " M i n d , M a t t e r , and Physicists." Foundations of

Physics. Vol. 2, n o . 2/3, 1972, p p . 129-147.
G a r e y , A . S . " S u p e r w e a k Interactions and the Biological Time D i r e c t i o n . "

Origins of Life. Vol. 9, 1978, p p . 1-5.
Gillespie, Neal C. Charles Darwin and the Problem of Creation. Chicago:

University of Chicago Press, 1979.
Glaessner, M.F. "Biological Events and the Precambrian Time Scale." In

Adventures in Earth History. E d i t e d by P. Cloud. San Francisco: W. H.
F r e e m a n and C o . , 1970.
G n e d e n k o , Boris V. Theory of Probability. 4th ed. Bronx, New York:

Chelsea, 1968.
Goldschmidt, Richard. The Material Basis of Evolution. New H a v e n : Yale

University Press, 1940.
G o s v ā m l , Satsvarūpa D ā s a . Readings in Vedic Literature. Los Angeles:

Bhaktivedanta Book Trust, 1977.
G o u l d , Stephen J. " C h a n c e R i c h e s . " Natural History, N o v e m b e r 1980,

p p . 36-44.
G o u l d , Stephen J. " H e n ' s T e e t h and H o r s e ' s T o e s . " Natural History, July

1980, p p . 2 4 - 2 8 .
G o u l d , S t e p h e n J. The Panda's Thumb. New York: W.W. N o r t o n and C o . ,

1980.
G o u l d , S t e p h e n J . a n d E l d r i d g e , Niles. " P u n c t u a t e d E q u i l i b r i a : T h e
T e m p o and M o d e of Evolution R e c o n s i d e r e d . " Paleobiology. Vol. 3,
1977, p p . 1 1 5 - 1 5 1 .
Grasse, Pierre-P. Evolution of Living Organisms. New York: Academic

Press, 1977.
H a d a m a r d , Jacques. The Psychology of Invention in the Mathematical

Field. Princeton: Princeton University Press, 1949.
H a r l a n d , W . B . , et. al., eds. The Fossil Record. London: Geological

Society of L o n d o n , 1967.
Bibliography 243
Hawking, S.W. and Israel, W., eds. General Relativity. Cambridge-

Cambridge University Press, 1979.
Heisenberg, Werner. Physics and Beyond. New York: H a r p e r and R o w ,

1971.
Heisenberg, Werner. " T h e R e p r e s e n t a t i o n of N a t u r e in C o n t e m p o r a r y

Physics." Daedalus. Vol. 87, n o . 3, 1958, pp. 95-108.
H e l m h o l t z , H e r m a n n Von. Ūber die Erhaltung der Kraft. Ostwald's

Klassiker der E x a k t e n Wissenschaften N r . 1, 1847.
H i n k e l , P e t e r C. and McCarty, Richard E. " H o w Cells M a k e A T P . "

Scientific American, March 1978, p p . 104-123.
Huffman, D . A . "A M e t h o d for the Construction of Minimum R e d u n

dancy C o d e s . " Proceedings of the I.R.E. Vol. 40, S e p t e m b e r 1952,
pp.1098-1101.
Huxley, Julian. Evolution in Action. New York: H a r p e r and Brothers,

1953.
Huxley, T h o m a s H. Essays on Some Controverted Questions. London:

Macmillan and C o . , 1892.
J a c o b , K.; J a c o b , C h i n n a ; and Shrivastava, R . N . Nature. Vol. 172,

n o . 4 3 6 9 , 1953, pp. 166-167.
J a m m e r , Max. The Philosophy of Quantum Mechanics. New York: J o h n

Wiley and Sons, 1974.
J a u c h , J . M . Are Quanta Real? Bloomington: Indiana University Press,

1973.
J o n e s , J a m e s P. "Recursive U n d e c i d a b i l i t y — A n Exposition." American

Mathematical Monthly, S e p t e m b e r 1974, p p . 724-738.
Kavirāja, Krsnadāsa G o s v ā m l . Śrī Caitanya-caritāmrta. Bengali text,

translation, and c o m m e n t a r y by A . C . Bhaktivedanta Swami Prabhu
pāda. Ādi-līlā 3 vols., Madhya-līlā 9 vols., Antya-līlā 5 vols. Los
Angeles: Bhaktivedanta Book Trust, 1974-1975.
K e p n e r , W . A . ; Gregory, W . C . ; and Porter, R . J . " T h e Manipulation of

the Nematocysts of Chlorohydra by M i c r o s t o m u m . " Zoologisher
Anzeiger. Vol. 121, J a n . - J u n e 1938, pp. 114-124.
Kleene, Stephen C. Introduction to Metamathematics. New York: Van

N o s t r a n d , 1952.
244 Bibliography
Kolmogorov, A n d r e i N. "Logical Basis for Information Theory and Proba-

bility T h e o r y . " IEEE Transactions on Information Theory. Vol. IT-14,
n o . 5, September 1968, p p . 662-664.
Layzer, David. " T h e A r r o w of T i m e . " The Astrophysical Journal.

Vol. 206, 1 J u n e 1976, p p . 559-569.
Lewin, R. "Evolutionary Theory U n d e r F i r e . " Science. Vol. 210,
21 N o v e m b e r 1980, p p . 883-887.
Macbeth, N o r m a n . Darwin Retried: An Appeal to Reason. Boston:

G a m b i t , 1971.
Mattuck, Richard D. and Walker, Evan H. "The Action of Consciousness

on Matter: A Q u a n t u m Mechanical Theory of Psychokinesis." In
The Iceland Papers. Edited by A. Puharich. Amherst, Wisconson:
Essentia Research Associates, 1979.
Mayr, Ernst. "The Emergence of Evolutionary Novelties." In Evolution

After Darwin. Vol. 1, The Evolution of Life. Edited by S. Tax. Chi-
cago: University of Chicago Press, 1960.
McDougall, I.; C o m p s t o n , W.; and H a w k e s , D . D . Nature. Vol. 198,

n o . 4880, 1963, p p . 564-567.
Messiah, A l b e r t . Quantum Mechanics. Vol. 2. Amsterdam: N o r t h -

Holland P u b . C o . , 1965.
M o n o d , Jacques. Chance and Necessity. New York: Alfred A. Knopf,

1971.
Mott, N . F . "The Wave Mechanics of oc-Ray Tracks." Proceedings of the
Royal Society, L o n d o n . Vol. 126, 1929, p p . 79-84.
Nagel, Ernest. The Structure of Science. New York and Burlingame:
H a r c o u r t , Brace, and World, 1961.
Newton's Principia. Translated by A n d r e w M o t t e . New York: Daniel

A d e e , 1846.
O p a r i n , A . I . The Origin of Life. 2d ed. New York: Dover, 1953.
Orgel, Leslie E. The Origins of Life. New York: Wiley, 1973.
Poincare, Henri. The Foundations of Science. Lancaster, Pa.: The Science
Press, 1946.
Popper, Karl R. and Eccles, John C. The Self and Its Brain. New York:
Springer-Verlag, 1977.
Bibliography 245
Prigogine, Ilya; Nicolis, G r e g o i r e ; and B a b l o y a n t z , A g n e s . " T h e r m o -

dynamics of Evolution." Physics Today, N o v e m b e r 1972, p p . 2 3 - 2 8 .
Rensberger, B. "Recent Studies Spark Revolution in Interpretation of
Evolution." The New York Times, 4 N o v e m b e r 1980, p. C 3 .
Rensch, Bernhard. Evolution Above the Species Level. New York:
Columbia University Press, 1960.
R o s e n f e l d , L e o n . " T h e M e a s u r i n g Process in Q u a n t u m M e c h a n i c s . "

Suppl. Progr. Theor. Phys., extra n. 1965, p p . 2 2 2 - 2 3 1 .
Ruelle, David. "Strange A t t r a c t o r s . " Mathematical Intelligencer. Vol. 2,

n o . 3 , 1 9 8 0 , p p . 126-137.
Russell, Bertrand. "A Free Man's Worship" (1903). In Mysticism and

Logic, and Other Essays. London: Allen and Unwin, 1963.
Satir, Peter. " H o w Cilia M o v e . " Scientific American, O c t o b e r 1974,

pp.44-52.
Schr dinger, Erwin. What is Life? and Mind and Matter. Cambridge:
Cambridge University Press, 1967.
Shannon, Claude E. "A Mathematical Theory of Communication." Bell

System Technical Journal. Vol. 27, July 1948, p p . 3 7 9 - 4 2 3 .
S h e r m a n , Irwin W. and S h e r m a n , Vilia G. Biology: A Human Approach.

New York: Oxford University Press, 1975.
Shklovskii, Iosif S. and Sagan, Carl. Intelligent Life in the Universe. San
Francisco: H o l d e n - D a y , 1966.
Simpson, George G. This View of Life. New York: H a r c o u r t , Brace, and

World, 1964.
Slater, John C. Quantum Theory of Molecules and Solids. Vol. 1. New

York: M c G r a w Hill, 1963.
Smith, John Maynard. "Hypercycles and the Origin of Life." Nature.

Vol. 280, 9 August 1979, p p . 445-446.
Smith, J o h n Maynard. " T h e Limitations of Evolutionary T h e o r y . " In The

Encyclopedia of Ignorance, edited by R. Duncan and M. Weston-
Smith. New York: Pocket Books, 1978.
Snelling, N . J . Nature. Vol. 198, n o . 4885,1963, p p . 1079-1080.

246 Bibliography
Staehelin, L. A n d r e w and Hull, Barbara E. "Junctions Between Living

Cells." Scientific American, May 1978, pp. 141-152.
Stainforth, R . M . Nature. Vol. 210, n o . 5033, 1966, p p . 292-294.
Tax, Sol and Callender, Charles, eds. Evolution After Darwin. Vol. 3,
Issues in Evolution. Chicago: University of Chicago Press, 1960.
T h a k u r , Bhakti Siddhanta Saraswati. Shri Brahma-Samhita. Madras,
India: Sree Gaudiya M a t h , 1958.
T h o m p s o n , Richard. "A Measure of Shared Information in Classes of
P a t t e r n s . " Pattern Recognition. Vol. 12, no. 6, D e c e m b e r 1980,
p p . 369-379.
Tolman, Richard C. The Principles of Statistical Mechanics. Oxford:

Claredon Press, 1938.
Valentine, James W. "The Evolution of Multicellular Plants and A n i m a l s . "
Scientific American, September, 1978, pp. 141-158.
Villee, Claude A. and Dethier, Vincent G. Biological Principles and

Processes. Philadelphia, Pa.: W . B . Saunders C o . , 1971.
von N e u m a n n , John. Mathematical Foundations of Quantum Mechanics.
Princeton: Princeton University Press, 1955.
von N e u m a n n , John. Theory of Self-Reproducing Automata. Edited by

A r t h u r Burkes. U r b a n a , Illinois: University of Illinois Press, 1966.
Wadia, Darashaw N. The Geology of India. 3d ed. L o n d o n : Macmillan,
1953.
Watson, James D. Molecular Biology of the Gene. 3d ed. Menlo Park,
California: W . A . Benjamin, 1977.
Weinberg, Steven. "Conceptual Foundations of the Unified Theory
of Weak and Electromagnetic Interactions." Science. Vol. 210,
12 D e c e m b e r 1980, p p . 1212-1218.
Weinberg, Steven. The First Three Minutes. New York: Bantum B o o k s ,

1977.
Weinberg, Steven. "The Forces of N a t u r e . " American Scientist. Vol. 65,

• M a r c h - A p r i l 1977, p p . 171-176.
Weisz, Paul B. Elements of Biology. New York: M c G r a w Hill Book C o . ,
1969.
Bibliography 247
W e i z e n b a u m , J o s e p h . Computer Power and Human Reason. San

Francisco: W.H. F r e e m a n and C o . , 1976.
Whitehead, Alfred N. Process and Reality: An Essay in Cosmology. Cor-

rected edition. Edited by D . R . Griffin and D . W . Sherburne. New
York: T h e Free Press, 1978.
Wigner, Eugene P. "Epistemological Perspective on Q u a n t u m T h e o r y . "
In Contemporary Research in the Foundations and Philosophy of
Quantum Theory. Edited b y C . A . H o o k e r . Boston: Reidel Publishing
C o . , 1973.
Wigner, Eugene P. " O n Hidden Variables and Q u a n t u m Mechanical

Probabilities." American Journal of Physics. Vol. 38, no. 8, August
1970, p p . 1 0 0 5 - 1 0 0 9 .
Wigner, Eugene P. "Physics and the Explanation of Life." Foundations of
Physics. Vol. 1, no. 1, 1970, p p . 3 5 - 4 5 .
Wigner, Eugene P. " R e m a r k s on the Mind-Body Q u e s t i o n . " In The Scien-

tist Speculates. Edited by I.J. Good. New York: Basic Books, 1962.
Wigner, Eugene P. "The Probability of the Existence of a Self Reproduc-

ing U n i t . " In The Logic of Personal Knowledge. London: Routledge
and Kegan Paul, 1961.
Wigner, Eugene P. "Two Kinds of Reality." The Monist. Vol. 48, 1964,
pp.248-264.
Wilson, Edward O. On Human Nature. Cambridge: Harvard University
Press, 1978.
Wilson, E. G r a n t . The Mystery of Physical Life. New York: Abelard
Schuman, 1964.
Winston, Patrick H. Artificial Intelligence. Reading, Mass.: Addison-

Wesley, 1977.
Winston, Patrick H. "The M I T R o b o t . " In Machine Intelligence 7. Edited
by B. Meltzer and D. Mitchie. New York: John Wiley and Sons, 1972.
Yockey, H u b e r t P. "A Calculation of the Probability of Spontaneous

Biogenesis by Information T h e o r y . " Journal of Theoretical Biology.
Vol. 67, 1977, p p . 377-398.
Yockey, Hubert P. " O n the Information Content of Cytochrome c."

Journal of Theoretical Biology. Vol. 67, 1977, pp. 345-376.
248 Bibliography
Z u k a v , Gary. The Dancing Wu Li Masters. New York: William Morrow

and C o . , 1979.
Index
A b s o l u t e c h a n c e , 148-153, 158. See B o u n d a r y conditions, 1 0 5 - 1 0 8 , 1 2 7 -

also C h a n c e ; Probability 129,132, 134, 234, 237
Absolute symbols, 25, 222-225 Brahma-samhitā, 80, 163
Aesthetics, s t a n d a r d s of, 89, 175 Brain, 14, 15; c o m p a r e d with
Algorithms, origin of complex, 119— c o m p u t e r , 36, 219; a n d conscious
120,176. See also C o m p u t e r s ; ness, 3 6 - 3 7 , 219; evolution of, 195;
Information c o n t e n t ; Information as distinct from material mind,
theory 219-220
A n g i o s p e r m s . See Flowering plants
A r r h e n i u s , Svante, 129 Caitanya M a h ā p r a b h u , 7, 224-225
Artificial intelligence, 28; a n d plan for Callixylon, logs of, 192
sentient c o m p u t e r , 33 C a m b r i a n period, 191-192
Ātmā. See Jīvātmā Canonical e n s e m b l e , 107. See also
Axelrod, D a n i e l , 190-192 Statistical mechanics
Central d o g m a , 113
Behavior: and consciousness, 33-34; Central processing unit ( C P U ) , 28-31
in animals, 119-120; in m a n , C e r e b r o s c o p e , 27
120-121 C h a i t i n , G r e g o r y , 136-137,227
Behavioral psychology, 22, 89 C h a n c e , 52, 73, 104, 110, 132;
B e t h e , H a n s , 84, n. 30 detailed discussion of, C h a p . 6;
Bhagavad-gītā, 7, 40, 161, 164, and inspiration, 172. See also
165,169, 177; a n d definition of Probability; Absolute chance
conscious self, 2 3 ; a n d science of C h u r c h ' s thesis, 30, 38
consciousness, 50, 7 5 - 7 6 , 8 1 ; on Cloud c h a m b e r , 52
mind-body interaction, 178-181; Codd,E.F.,237
on p u r p o s e of material world, Cognitive engineering. See Artificial
204-205; n o n m e c h a n i s t i c world intelligence
view of, 212-225 C o m p l e m e n t a r i t y , 70
Bhāgavata Purāna, 7, 218, 221, 222 C o m p u t e r : brain c o m p a r e d with, 15;
B h a k t i v e d a n t a Swami P r a b h u p ā d a , description of m o d e r n digital, 2 8 -
A.C, 7 3 1 ; p r o g r a m , example of, 2 9 - 3 0 ;
Bhakti-yoga, 7, 25; as science, 9, 10, levels of organization in, 3 1 - 3 2 ;
214-225 and information t h e o r y , 227-228
Big bang theory, 99, 106-107 Conscious self. See Jīvātmā
Biology: mechanistic premises of, 5, Consciousness, 8; not explainable by
13; a n d views of biologists, 50 physics a n d chemistry, 13; as
B o h r , Niels, 6 1 ; and c o m p l e m e n subjective, 15-16, 25; recognized
tarity, 70 as nonphysical by scientists, 22;
B o n d s , chemical, 236 behavioral criterion for, 33; as
249
250 Index
Consciousness Escherichia coli

objectively real, 3 4 - 3 5 ; its relation complex metabolic system of, 118;
with states of a m a c h i n e , 3 5 - 3 6 ; molecular m o t o r s of, 196
and the brain, 3 6 - 3 7 , 40, 2 1 9 - 2 2 0 ; Equality, absolute basis for political, 24
science of, n e e d e d , 4 9 - 5 0 ; not ex- Evil, and divine creation, 203,
plainable by q u a n t u m mechanics, 204-205
62; and Darwinian evolution, 90; Evolution: D a r w i n i a n , 2, 88, 90, 92,
as unified, 160; universal, 161— 103,124,135-136,184-186;
1 6 4 , 1 7 7 ; and c o m p u t e r o p e r a - molecular, 2, 103-105; e m e r g e n t ,
tions, 166, n. 12; and inspiration, 8 7 - 8 8 , 9 1 - 9 3 ; and information
177; nonmechanistic n a t u r e of, theory, C h a p . 5; and absolute
212; purification of, 2 1 4 - 2 2 5 chance, C h a p . 6; as process of
Constants: in physics, 25; transcen- probability zero, 132-133, 151,
d e n t a l , 2 5 , 2 2 2 - 2 2 5 ; governing 153; as invisible, 186; and inter-
evolution, 129 mediate forms, 192-202; and
C r e a t i o n , 184, 2 0 0 - 2 0 2 , 203 theology, 2 0 2 - 2 0 7 ; three prin-
ciples of, 235
Darwin, Charles, 124,156,184-185, Explanation, of complex form is not
1 9 5 , 1 9 9 , 203, 205 possible, 1 3 4 , 1 3 8 , 2 1 1 - 2 1 2
D a t a compression, 152; physical Extrapolation, unjustifiable, 5 - 6 ,
theories as examples of, 159; 80-81
and the unity of conscious
perception, 160 Faith, 80, 2 1 7 - 2 1 8
D e s c a r t e s , R e n e , 44 Fisher, R o n a l d , 104
Design, 2 0 2 , 2 0 4 , 206 Flowering plants, origin of, 186-192
Devotional service, 215, 2 2 1 , 2 2 2 Fossil record, 184-192
D i r a c , P. A . M . , 237, n. 2 Functionalism, 38
D N A , 112,130,198
D o b z h a n s k y , T h e o d o s i u s , 132, Galois, Evariste, 175
153-156 G a u s s , Karl, 170
D o g m a . See Central dogma G e n e r a t i n g function, 114, 116,
230-232
Eccles, J o h n , 87, 9 1 , 9 4 G e n e s : regulative, 198-202;
Eigen, Manfred, 130-131 structural, 198
Einsten, A l b e r t , 9 9 , 1 4 3 , 1 6 9 - 1 7 0 , Geological time scale, 187-188
211,225 G l a s h o w , Sheldon, 9 9 , 1 4 3
Eldridge, Niles, 185 G o d . See Supreme Person; Krsna
Empiricism, limits of, 135, 137-138 Goldschmidt, Richard, 197,198
E n z y m e s . See Proteins G o u l d , S t e p h e n , J., 132, 185, 203
E p i p h e n o m e n a l i s m , 40, 44 G o v i n d a . See Krsna
E p i p h e n o m e n o n , 22, 51 Guru, 2 1 6 - 2 1 7 , 218, 221. See also
Epistemology: of q u a n t u m Teacher
mechanics, Sections 3 . 1 - 3 . 3 ; of
bhakti-yoga, C h a p . 9 Hadamard, Jacques, 171,174-175
Escherichia coli, 18, 111, 131; genetic H a m i l t o n i a n , 1 0 1 - 1 0 3 , 1 0 7 , 233
information content of, 112; Helmholtz, H e r m a n n Von, 101
Index 25
H i d d e n variables, 70 Karma, 1 8 0 - 1 8 1 , 2 1 4
Histories, q u a n t u m mechanical, Kirtanam, 224
66-68 Kolmogorof, A . I . , 227
H o r s e s : information c o n t e n t of, 127; Krsna, 79, 80, 8 2 , 1 6 1 , 163, 164,
toes of, 198, 200-201 178-179, 181, 204, 212, 214-218,
Hrsikeśa, 161. See also K r s n a 221-225. See also S u p r e m e Person
Huxley, Julian, 6, 183
Huxley, T h o m a s , 22 Laplace, Pierre Simon d e , 143, 158
H y p o t h e s e s , 169-170 Laws of n a t u r e . See N a t u r a l laws
Laws of physics. See N a t u r a l laws
Identity t h e o r y , 39 Layzer, David, 107
Infinite regress: in explanation of Life: as chemical process, 13, 100;
conscious self, 42; in e x p l a n a t i o n of as fundamental principle of n a t u r e ,
origin of life, 98 98; detailed discussion of origin
Infinity, 162 of, C h a p . 5; i m p r o b a b l e as a result
Information c o n t e n t , 97; defined, 105, of physical processes, 122-123. See
227; n o t a t i o n for, 109, 227; u p p e r also Evolution; Consciousness;
b o u n d s o n , for p r o t e i n s , 112-113; Jīvātmā
lower b o u n d s o n , for genetic
coding, 114-116; and evolution, M a c r o m u t a t i o n s , 197, 198. See also
122-129; S h a n n o n ' s formulation Mutations
of, 152; and r a n d o m n e s s , 152, 211; M ā d h a v a , 161. See also Krsna
of universal t h e o r y , 159; of uni Madhva, 7
verse, 161; s h a r e d , 229; of physio- Mahāmantra. See Mantra
chemical m o d e l , 233-237 M a m m a l s : D N A in genes of, 112;
Information theory: general discus information c o n t e n t of p r o t e i n s in,
sion of, C h a p . 5; a n d c h a n c e , 116; evolution of, ruled o u t , 127
152-153; technical discussion of. Mantra, 25, 224
A p p e n d i c e s 1 and 2 M a t h e m a t i c a l proof, 170, 173-174
Initial conditions, 105-108, 127-129, M a y r , Ernst, 193
132, 134, 237; as script, 158 M e a n i n g , 74, 89, 165. See also
Inspiration, C h a p . 7 Purpose
Intelligence: artificial, 28, 3 3 ; M e a s u r e m e n t p r o b l e m , in q u a n t u m
complexity of h u m a n , 120; mechanics, 6 0 - 6 1
origin of, 176; a n d design of bodily Mechanistic world view, 27; m e a n i n g
plans, 201-202; evolution of, 236 not definable in, 7 4 , 1 6 4 - 1 6 5 ;
Interactionism, 40, 44-45 limits us to correlating m e a s u r a b l e
Iśopanisad, 162 quantities, 77; defined, 143; c a n n o t
e n c o m p a s s conscious awareness,
Jīvātmā, 8; axioms for, 2 3 - 2 4 ; senses 160; c a n n o t be complete and
of, 2 4 - 2 5 , 4 1 - 4 2 , 213-214, 217, unified, 162-163; excludes
220, 223-224; investigation of 44, p u r p o s e , 204
7 7 - 7 8 ; size of, 45; and its inter M e m o r y , of c o m p u t e r , 2 8 - 2 9
action with the body, 177-181; and M e t a b o l i s m , 118
its relationship with K r s n a , Meta-laws, and emergent evolution, 92
204-205,212-225 Microstomum, 193-195, 197
252 Index
M i n d : distinguished from body, Objectivity, of observation, 7 8 - 7 9 ,

C h a p . 4; distinguished from body 217-218
a n d conscious self, 219-221 Observable, quantum mechanical, 108
Mind-body p r o b l e m , 3 8 - 4 5 , 8 7 - 9 1 , O b s e r v a t i o n function, 125-127, 229
177-181,219-220 O c c a m ' s razor, 90
M I T r o b o t , 120 O r g a n s , complexity of, 119
M o d e l , of physical system, 121-122,
233-237 P a r a m ā t m ā , 78, 169. See also
Molecules: biological, 111-116; Supersoul
r e p r e s e n t a t i o n of, 236-237 Parapsychology, 75, 219
Monod, Jacques, 91,133 Personality: a b s o l u t e , 4, 6, 161-165, '
M o n s t e r , hopeful, 197-202 212-225; complexity of, 120; a n d
Morality, 3, 89. See also Values evolution, 130
M o z a r t , Wolfgang, 171,176 Physics, classical, 1 0 0 - 1 0 1 , 1 5 6 - 1 5 8 .
Multiple universe theory, 7 2 - 7 4 See also Q u a n t u m mechanics;
Mundaka Upanisad, 45 N a t u r a l laws; Statistical
Mutations, 97,110,125,183, mechanics; Theories
192-193, 197-202; and p r o b a P o i n c a r e , H e n r i , 171-175,180
bility, 200, 235 Pollen, fossil, 191
P o p p e r , Karl, C h a p . 4
N a t u r a l selection, 9 7 - 9 8 , 104, 156, Positivism, 74, 77, 141, n. 42
1 7 6 , 1 8 3 , 1 9 2 - 1 9 3 , 202, 235, 236; P r e - C a m b r i a n organisms, 191
implicit in physical m o d e l , 110; Prigogine, Ilya, 130
c a n n o t select complex form, 124, Primordial s o u p , 105
133; wand of, 145 Probability: of molecule, 108, 121; in
N a t u r a l laws, 9 7 , 1 3 4 - 1 3 5 ; violated by physical m o d e l , 121-122; of
nonphysical entity, 45, 90; in evolution, 122,127; and infor
q u a n t u m mechanics, 5 5 - 7 2 ; a n d m a t i o n c o n t e n t , 122,126-128,
fundamental forces, 77; hier 227-229; and statistical laws, 146;
archical organization of higher, 82, of m u t a t i o n s , 200
178-179, 180; cannot g e n e r a t e life P r o g r a m m i n g language, for use in
if simple, 104, information c o n t e n t information theory, 227-228,
of, 122, A p p e n d i x 2; of high 233-234
complexity, 127; as source of Proteins: myoglobin, 18, 19; of
information, 128,132; as E. coli, 111; of v e r t e b r a t e s , 113;
conventions of t h e S u p r e m e , 164, function of, 115-116,130;
178,180; of physiochemical m o d e l , information c o n t e n t of, 116
233-234 Psilophytes, 191-192
N e m a t o c y s t s , 193-195 Psychophysical parallelism, 39
Nerve cells, 13-17 Punctuated equilibrium, 185-186,188
N e w t o n , Isaac, 101 P u r p o s e , 2, 74, 89; of material world,
Nonmechanistic world view: outlined, 204; of conscious self, 214. See also
7-10, 23, 38, 7 6 - 8 2 , 9 3 , 1 6 1 - 1 6 5 , Meaning
177-181, 202; m e t h o d of investi
gating, C h a p . 9 Q u a n t u m field theory, 99. See also
Numerical describability, 24 Renormalization
Index 253
Q u a n t u m mechanics, 19, 45; exten Slater, J o h n , 100

sive discussion of, Sections 3 . 1 - Smith, J o h n M a y n a r d , 1
3.3; and biological p r o b l e m s , Solipsism, 16, 62
5 1 - 5 2 , 72, 235; and classical Soul. See Jīvātmā
physics, 6 1 ; and t h e observer, Spin, q u a n t u m mechanical, 234
6 1 - 6 3 ; and reaction-diffusion Spinoza, 39
models, 131; and absolute c h a n c e , Sravanam, 222
148; a n d information t h e o r y , State vectors: example of, 5 6 - 6 0 ;
233-234; relativistic t h e o r y , 234 m a t h e m a t i c a l details concerning,
8 3 - 8 4 , n. 24; a n d molecular
Radical materialism, 88 descriptions, 236
Radioactivity: q u a n t u m mechanical Śrimad-Bhāgavatam. See Bhāgavata
description of, 5 6 - 6 0 ; as example Purāna
of absolute c h a n c e , 148-149 Statistical laws, 146-148
Rāmānuja, 7 Statistical mechanics, 107, 121, 128
Reaction-diffusion e q u a t i o n s , 72, Statocyst, 195-196
130-131 Subliminal self, 171-172, 173
R e d u c t i o n of t h e wave function, 64, Superconsciousness, 78-79
68-69 Supersoul, 179,180,216, 219,220. See
R e d u c t i o n i s m , and consciousness, also Paramātmā
13-21 S u p r e m e P e r s o n , 8, 7 9 - 8 0 , 161-165,
Relativity, general, 73, 99 178-181, 204, 212-225. See also
Religion, and its conflict with Krsna
science, 3, 4 Symbolic description, 117-118; of
R e n o r m a l i z a t i o n , 237, n. 2 h u m a n personality, 120-121; a n d
R i e m a n n , B e r n h a r d , 174 observation, 125-127
Russell, B e r t r a n d , 2 Symbols. See Absolute symbols
Sagan, Carl, 132

T e a c h e r , spiritual, 79
Salam, A b d u s , 99, 143
Theology: n a t u r a l , 135-136, 138,
Schrodinger, Erwin, 163-164
205-206, 213; negative, 202-207
Schr0dinger e q u a t i o n , 55, 56, 64, 65,
T h e o r i e s : basis for discussing, 8; to
67, 7 3 , 102, 103, 127; information
explain consciousness, 3 8 - 4 1 ;
c o n t e n t of, 105, 121, 233-234
q u a n t u m mechanical, C h a p . 3 ;
Sea slug, 197
mechanistic, must b e i n c o m p l e t e ,
Self, as real entity, 9 3 - 9 4 . See also
156, 159, 162; deterministic,
Jīvātmā
157-158; to explain inspiration,
Self-reproduction: of a u t o m a t a ,
172-173; of physics underlying
110, 235-236; and q u a n t u m
chemistry, 233-234
mechanics, 235
Tunneling, q u a n t u m mechanical, 53
Senses: of jīvātmā, 2 4 - 2 5 , 4 1 - 4 2 ,
Turing, A l a n , 172
213-214,217,220,223,224;
Turing m a c h i n e , 110
material, 218-220, 221
Set t h e o r y , 162
S h a n n o n , C l a u d e , 152 Undecidability, 142, n. 48
Simpson, G e o r g e G . , 132-133 Unified field t h e o r y , 143
254 Index
Unity: in physics, 143,156, 159-160, W a t s o n , J a m e s , 100, 111, 113

2 1 1 ; nonmechanistic, 161-165 Wave-particle duality, 5 4 - 5 5 , 71
Weinberg, Steven, 2 , 9 9 , 1 4 3 , 1 4 9
Vaisnavism, 7 W h a l e s , evolution of, 201-202
Values, 3, 93 W h i t e h e a d , Alfred N . , 39, 40
Vitalism, 5 , 6 , 5 0 Wigner, E u g e n e , 22, 4 5 - 4 6
Voidism, of m o d e r n scientific world World t h r e e , 93
view, 74 World view. See Mechanistic world
von N e u m a n n , J o h n , 39; and q u a n t u m view; Nonmechanistic world view
mechanical observation, 6 1 ;
density matrix of, 6 4 , 1 0 7 - 1 0 8 ; Yockey, H u b e r t , 131
cellular a u t o m a t o n model of, 110,
235-236

Mechanistic and Nonmechanistic Science PDF

Uploaded by

Copyright:

Available Formats

Mechanistic and Nonmechanistic Science PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mechanistic and Nonmechanistic Science PDF

Uploaded by

Copyright:

Available Formats

MECHANISTIC

Library of Congress Cataloguing in Publication Data

Bala Books, 39 Dawes Avenue, Lynbrook, N . Y . 11563

His Divine Grace

Chapter 1. Searching Past the Mechanics of Perception . . . . 13

Chapter 3. Dialogue on Consciousness and the Quantum . . . . 49

Chapter 4. Karl Popper on the Mind-Body Problem—

PART II. FORM

Chapter 5. Information Theory and the Self-organization

Chapter 7. On Inspiration 169

Chapter 8. The Doctrine of Evolution 183

PART III. CONCLUSION

Chapter 9. The Epistemology of Transcendental

Appendix 1. * A Discussion of Information Theory 227

(1) Matter can be represented by numbers that correspond directly or

The world view of modern physics is essentially mathematical, and to a per-

The mechanistic world view also exerts a powerful influence on human

rigid, unverifiable doctrines, many of which have been discredited by mod-

sonality is built into the nature of things, it is not possible to formulate a

is illustrated by the remark, written in a standard biology textbook, that

tics of a scientific theory. It should provide a logically consistent descrip­

My discussion of modern scientific theories rests entirely on logic, the

subject of artificial intelligence, the classical theories of the mind-body

only way of giving a tangible demonstration of the validity of these princi­

1. Smith, "The Limitations of Evolutionary Theory," p. 239.

turn evoke corresponding sequences of muscular contractions. These or-

way: The reality of my own conscious perceptions is certainly undeniable,

Figure 3. A magnified image of a single nerve cell, showing various kinds of

molecular interactions. Figure 4 depicts the three-dimensional structure of

Figure 4. The three-dimensional structure of myoglobin, a large biological

the nature of consciousness than any of the higher-order descriptions we

the spatial distribution of electrons within an organic molecule. Those who

Figure 6. A quantum mechanical diagram describing the electronic structure of

4A00 DD21004B 00130 4A11 DD7E02 00180 4A20 DD7701 00230

Figure 7. The reduction of mathematical equations to fundamental rules for symbol

senses. Thus we began our investigation by describing a man with symbols

sciousness itself. When contemplating records of measurements made on

cannot possibly learn more about consciousness than we already know.

1. The jīvātmā is the elementary unit of consciousness.

The first of these axioms introduces the self as an irreducible, natural

Hare Krsna, Hare Krsna, Krsna Krsna, Hare Hare

According to the system of bhakti-yoga, such chanting can purify the

2.1 How a Computer Works

they do not correspond exactly to those of any particular computer).

At this point, let us consider our hypothetical sentient computer. Ac-

2.2 Artificial Intelligence and Hierarchies of Function

of the combination of elementary operations that is actually used to com-

1. Find the square root of X.

Figure 2. Levels of organization of the program in Figure 1. The program in Figure

the field of artificial intelligence generally visualize their projected "intel-

2.3 Subjective Consciousness in Machines and Humans

Figure 3. The relation between

To clearly distinguish this conception of consciousness from the behavioral

(1) This entity must possess sufficient powers of discrimination to

abstract properties undergo as time passes and the arrangements

Figure 4, The relation between

by statements ( 1 ) and ( 2 ) must exist to account for the connection between

depend on some superficial details of our understanding of the brain. On a

2.4 Several Nonmechanistic Theories

tics of a scientific theory. It should provide a logically consistent descrip

only way of giving a tangible demonstration of the validity of these princi

awkwardly into a mechanistic world picture as an inexplicable, incompa

fairly nontechnical way if we accept one basic mathematical transfor

vantage of the soul's particular mode of interaction. That is reason