[Dr. Gerald Edelman Ph.D.] Wider Than the Sky the(BookFi.org)

G E R A L D M . E D E L M A N

the phenomenal gift of consciousness

y a l e n o t a b e n e

yale university press new haven and london

First published in 2004 by Yale University Press.This Nota Bene edition published 2005 by

Yale University Press.

Copyright © 2004 by Gerald M. Edelman.

This book may not be reproduced, in whole orin part, including illustrations, in any form

(beyond that copying permitted by Sections 107 and108 of the U.S. Copyright Law and except by

reviewers for the public press) without written permissionfrom the publishers.

Printed in the United States of America.

For information about this and other Yale UniversityPress publications, please contact:U.S. office [email protected] office [email protected]

The Library of Congress has cataloged the hardcoveredition as follows:

Edelman, Gerald M.Wider than the sky : the phenomenal gift of

consciousness / Gerald M. Edelman.p. cm.

Includes bibliographical references and index.ISBN 0-300-10229-1 (alk. paper)1. Consciousness—Physiological aspects. 2.

Brain—Physiology. I. Title.QP411 .E346 2004153—dc22

ISBN 0-300-10761-7 (pbk.)

A catalogue record for this book is available fromthe British Library.

10 9 8 7 6 5 4 3 2 1

For Maxine

The Brain—is wider than the Sky—

For—put them side by side—

The one the other will contain

With ease—and you—beside—

The Brain is deeper than the sea—

For—hold them—Blue to Blue—

The one the other will absorb—

As sponges—Buckets—do—

The Brain is just the weight of God—

For—Heft them—Pound for Pound—

And they will differ—if they do—

As Syllable from Sound—

Emily Dickinson, c. 1862

Contents

Preface xi

Acknowledgments xv

1. The Mind of Man: Completing Darwin’sProgram 1

2. Consciousness: The Remembered Present 4

3. Elements of the Brain 14

4. Neural Darwinism: A Global Brain Theory 32

5. The Mechanisms of Consciousness 48

6. Wider Than the Sky: Qualia, Unity, andComplexity 60

7. Consciousness and Causation: ThePhenomenal Transform 76

8. The Conscious and the Nonconscious:Automaticity and Attention 87

9. Higher-Order Consciousness andRepresentation 97

10. Theory and the Properties of Consciousness 113

11. Identity: The Self, Mortality, and Value 131

12. Mind and Body: Some Consequences 140

Glossary 149

Bibliographic Note 181

Index 187

c o n t e n t sx

Preface

Consciousness is the guarantor of all we hold tobe human and precious. Its permanent loss is consideredequivalent to death, even if the body persists in its vitalsigns. No wonder, then, that consciousness has attractedspeculation and study across the ages. Over the pasttwenty-five years, I have written a number of books andpapers on the subject. My conviction that consciousnessis susceptible to scientific study has been supported bya sharp increase in the number of publications and sci-entific meetings on the subject.

These developments have prompted me to presentan account of consciousness to the general reader. Incarrying out this project, my goals were clear: to defineconsciousness and to offer as simple a view of the subjectas is consistent with clarity. The subject is a challengingone and it will certainly require a concentrated efforton the part of the reader. I can only promise that thereward for such effort will be a deeper insight into issuesthat are at the center of human concern. Accordingly,except when absolutely necessary, I have deliberatelyomitted many scholarly references, which may be found

in abundance in my previous works. Those interestedin further reading can find a number of excellent worksthat have informed this book listed in the BibliographicNote. I am aware that a great barrier to understandingscientific presentations rests in the inevitable use of tech-nical terms. The problem is compounded when oneconsiders details related to the brain and consciousness.For this reason, I have added a glossary that I hope willprovide some alleviation.

William James, whose descriptions of conscious-ness still stand as a high-water mark in the field, said:

Something definite happens when to a certainbrain-state a certain “sciousness” corresponds. Agenuine glimpse into what it is would be the sci-entific achievement, before which all past achieve-ments would pale. But at the present, psychologyis in the condition of physics before Galileo andthe laws of motion, of chemistry before Lavoisierand the notion that mass is preserved in all reac-tions. The Galileo and the Lavoisier of psychologywill be famous men indeed when they come, ascome they some day surely will, or past successesare no index to the future. When they do come,however, the necessities of the case will makethem “metaphysical.” Meanwhile the best way inwhich we can facilitate their advent is to under-stand how great is the darkness in which wegrope, and never forget that the natural-science

p r e f a c exii

assumptions with which we started are provisionaland revisable things.

I have puzzled over what James had in mind instating that successful scientific efforts to glimpse thebases of consciousness would necessarily be metaphysical.In any event, in this book I have tried to avoid extensivediscussion of metaphysical matters. I intend to deal withexplanations that rest solely on a scientific base. Myhope is to disenthrall those who believe the subject isexclusively metaphysical or necessarily mysterious.

A scientific analysis of consciousness must answerthe question: How can the firing of neurons give rise tosubjective sensations, thoughts, and emotions? To some,the two domains are so disparate as to be irreconcilable.A scientific explanation must provide a causal accountof the connection between these domains so that prop-erties in one domain may be understood in terms ofevents in the other. This is the task I have set myself inthis small book.

The title of the book comes from a poem by EmilyDickinson that appears as the epigraph. This poem waswritten in around 1862, before modern brain sciencebegan toward the end of the nineteenth century. I findit impressive that, in extolling the width and depth ofthe mind, Dickinson referred exclusively to the brain.As for my subtitle, it is a play on the remarkable natureof consciousness, as well as on its rendering to us of thesignals of this world.

p r e f a c exiii

Acknowledgments

I thank Drs. Kathryn Crossin, David Edelman, Jo-seph Gally, Ralph Greenspan, and George Reeke fortheir valuable criticism and useful suggestions. The illus-trations were composed by Eric Edelman; I am gratefulfor his skilled and patient responses to my sometimesidiosyncratic suggestions. Darcie Plunkett provided ex-cellent help in preparing this manuscript.

The Mind of ManC O M P L E T I N G D A R W I N ’ S P R O G R A M

n 1869, Charles Darwin found himself vexedwith his friend Alfred Wallace, the co-founder of thetheory of evolution. They had differed on several issuesrelated to that theory. But the main reason for Darwin’sdisturbance was a publication by Wallace concerningthe origin of the brain and mind of man. Wallace, whoby that time had spiritualist leanings, concluded thatnatural selection could not account for the human mindand brain.

Darwin wrote to him before publication: “I hopeyou have not murdered too completely your own and

1

my child,” meaning, of course, natural selection. Wal-lace, in fact, concluded that natural selection could notexplain the origin of our higher intellectual and moralfaculties. He claimed that savages and prehistoric hu-mans had brains almost as large as those of Englishmenbut, in adapting to an environment that did not requireabstract thought, they had no use for such structuresand therefore their brains could not have resulted fromnatural selection. Unlike Wallace, Darwin understoodthat such an adaptationist view, resting only on naturalselection, was not cogent. He understood that propertiesand attributes not necessarily needed at one time couldnevertheless be incorporated during the selection ofother evolutionary traits. Moreover, he did not believethat mental faculties were independent of one another.As he explained in his book The Descent of Man, forexample, the development of language might have con-tributed to the process of brain development.

This rich work has prevailed, along with Darwin’sother views, but the program he established remains tobe completed. One of the key tasks in completing thatprogram is to develop a view of consciousness as a prod-uct of evolution rather than as a Cartesian substance, orres cogitans, a substance not accessible to scientific analy-sis. A major goal of this book is to develop such a view.

What is required to carry out such a project? Beforeanswering this question, let us consider Darwin’s entryin his notebook of 1838: “Origin of man now proved—metaphysic must flourish—He who understands ba-

t h e m i n d o f m a n2

boon will do more towards metaphysics than Locke.”These statements point in the direction we must follow.We must have a biological theory of consciousness andprovide supporting evidence for that theory. The theorymust show how the neural bases for consciousness couldhave arisen during evolution and how consciousness de-velops in certain animals.

Two subtle but important issues strongly influenceour interpretation of these requirements. The first ofthese is the question of the causal status of conscious-ness. Some take the view that consciousness is a mereepiphenomenon with no material consequences. A con-trary view is that consciousness is efficacious—that itcauses things to happen. We will take the position,which we shall explore in detail later, that it suffices toshow that the neural bases of consciousness, not con-sciousness itself, can cause things to happen. The secondmajor challenge to any scientific account of conscious-ness is to show how a neural mechanism entails a subjec-tive conscious state, or quale, as it is called. Before wecan meet these two challenges, it is necessary to providea sketch of the properties of consciousness and considersome matters of brain structure and function.

t h e m i n d o f m a n3

ConsciousnessT H E R E M E M B E R E D P R E S E N T

e all know what consciousness is : it is whatyou lose when you fall into a deep dreamless sleep andwhat you regain when you wake up. But this glib state-ment does not leave us in a comfortable position to ex-amine consciousness scientifically. For that we need toexplore the salient properties of consciousness in moredetail, as William James did in his Principles of Psychol-ogy. Before doing so, it will help to clarify the subjectif we first point out that consciousness is utterly depen-dent on the brain. The Greeks and others believed that

4

consciousness resided in the heart, an idea that survivesin many of our common metaphors. There is now a vastamount of empirical evidence to support the idea thatconsciousness emerges from the organization and opera-tion of the brain. When brain function is curtailed—in deep anesthesia, after certain forms of brain trauma,after strokes, and in certain limited phases of sleep—consciousness is not present. There is no return of thefunctions of the body and brain after death, and post-mortem experience is simply not possible. Even duringlife there is no scientific evidence for a free-floating spiritor consciousness outside the body: consciousness is em-bodied. The question then becomes: What features ofthe body and brain are necessary and sufficient for con-sciousness to appear? We can best answer that questionby specifying how the properties of conscious experiencecan emerge from properties of the brain.

Before taking up the properties of consciousness inthis chapter, we must address another consequence ofembodiment. This concerns the private or personal na-ture of each person’s conscious experience. Here isJames on the subject :

In this room—this lecture room, say—there area multitude of thoughts, yours and mine, someof which cohere mutually, and some not. Theyare as little each-for-itself and reciprocally inde-pendent as they are all-belonging-together. Theyare neither: no one of them is separate, but each

c o n s c i o u s n e s s5

belongs with certain others and with none beside.My thought belongs with my other thoughts andyour thought with your other thoughts. Whetheranywhere in the room there be a mere thought,which is nobody’s thought, we have no means ofascertaining, for we have no experience of the like.The only states of consciousness that we naturallydeal with are found in personal consciousness,minds, selves, concrete particular I’s and you’s.

There is no mystery here. Since consciousness ar-rives as a result of each individual’s brain and bodilyfunctions, there can be no direct or collective sharingof that individual’s unique and historical conscious ex-perience. But this does not mean that it is impossibleto isolate the salient features of that experience by obser-vation, experiment, and report.

What is the most important statement one canmake about consciousness from this point of view? It isthat consciousness is a process, not a thing. James madethis point trenchantly in his essay “Does ConsciousnessExist?” To this day, many category errors have beenmade as a result of ignoring this point. For example,there are accounts that attribute consciousness specifi-cally to nerve cells (or “consciousness neurons”) or toparticular layers of the cortical mantle of the brain. Theevidence, as we shall see, reveals that the process of con-sciousness is a dynamic accomplishment of the distrib-uted activities of populations of neurons in many differ-


ent areas of the brain. That an area may be essential ornecessary for consciousness does not mean it is suffi-cient. Furthermore, a given neuron may contribute toconscious activity at one moment and not at the next.

There are a number of other important aspects ofconsciousness as a process that may be called Jamesianproperties. James pointed out that consciousness occursonly in the individual (that is, it is private or subjective),that it appears to be continuous, albeit continuallychanging, that it has intentionality (a term referring tothe fact that, generally, it is about things), and that itdoes not exhaust all aspects of the things or events towhich it refers. This last property has a connection tothe important matter of attention. Attention, particu-larly focal attention, modulates conscious states and di-rects them to some extent, but it is not the same as con-sciousness. I will return to this issue in later chapters.

One outstanding property is that consciousness isunitary or integrated, at least in healthy individuals.When I consider my conscious state at the time of thiswriting, it appears to be all of a piece. While I am payingattention to the act of writing, I am aware of a ray ofsunlight, of a humming sound across the street, of asmall discomfort in my legs at the edge of the chair, andeven of a “fringe,” as James called it, that is of objectsand events barely sensed. It is usually not entirely possi-ble to reduce this integrated scene to just one thing,say my pencil. Yet this unitary scene will change anddifferentiate according to outside stimuli or inner


thoughts to yet another scene. The number of such dif-ferentiated scenes seems endless, yet each is unitary. Thescene is not just wider than the sky, it can contain manydisparate elements—sensations, perceptions, images,memories, thoughts, emotions, aches, pains, vague feel-ings, and so on. Looked at from the inside, conscious-ness seems continually to change, yet at each momentit is all of piece—what I have called “the rememberedpresent”—reflecting the fact that all my past experienceis engaged in forming my integrated awareness of thissingle moment.

This integrated yet differentiated state looks en-tirely different to an outside observer, who possesses hisor her own such states. If an outside observer testswhether I can consciously carry out more than two taskssimultaneously, he will find that my performance deteri-orates. This apparent limitation of conscious capability,which is in contrast to the vast range of different innerconscious states, deserves analysis. I will consider its ori-gins when I discuss the difference between consciousand nonconscious activity.

So far, I have not mentioned a property that is cer-tainly obvious to all humans who are conscious. We areconscious of being conscious. (Indeed, it is just such aform of consciousness that impels the writing of thisbook.) We have scant evidence that other animals pos-sess this ability; only higher primates show signs of it.In the face of this fact, I believe that we need to makea distinction between primary consciousness and higher-


order consciousness. Primary consciousness is the stateof being mentally aware of things in the world, of havingmental images in the present. It is possessed not onlyby humans but also by animals lacking semantic or lin-guistic capabilities whose brain organization is neverthe-less similar to ours. Primary consciousness is not accom-panied by any sense of a socially defined self with aconcept of a past or a future. It exists primarily in theremembered present. In contrast, higher-order con-sciousness involves the ability to be conscious of beingconscious, and it allows the recognition by a thinkingsubject of his or her own acts and affections. It is accom-panied by the ability in the waking state explicitly to re-create past episodes and to form future intentions. At aminimal level, it requires semantic ability, that is, theassignment of meaning to a symbol. In its most devel-oped form, it requires linguistic ability, that is, the mas-tery of a whole system of symbols and a grammar.Higher primates, to some minimal degree, are assumedto have it, and in its most developed form it is distinctiveof humans. Both cases require an internal ability to dealwith tokens or symbols. In any event, an animal withhigher-order consciousness necessarily must also possessprimary consciousness.

There are different levels of consciousness. In rapideye movement (REM) sleep, for example, dreams areconscious states. In contrast with individuals in the wak-ing state, however, the dreaming individual is often gull-ible, is generally not conscious of being conscious, is not


connected to sensory input, and is not capable of motoroutput. In deep or slow-wave sleep, short dreamlike epi-sodes may occur, but for long periods there is no evi-dence of consciousness. In awaking from the uncon-sciousness induced by trauma or anesthesia, there maybe confusion and disorientation. And, of course, theremay be diseases of consciousness, such as schizophrenia,in which hallucinations, delusions, and disorientationcan occur.

In the normal conscious state, individuals experi-ence qualia. The term “quale” refers to the particularexperience of some property—of greenness, for in-stance, or warmth, or painfulness. Much has been madeof the need for providing a theoretical description thatwill allow us directly to comprehend qualia as experi-ences. But given that only a being with an individualbody and brain can experience qualia, this kind of de-scription is not possible. Qualia are high-order discrimi-nations that constitute consciousness. It is essential tounderstand that differences in qualia are based on differ-ences in the wiring and activity of parts of the nervoussystem. It is also valuable to understand that qualia arealways experienced as parts of the unitary and integratedconscious scene. Indeed, all conscious events involve acomplex of qualia. In general, it is not possible to experi-ence only a single quale—“red,” say—in isolation.

I shall elaborate later on the statement that qualiareflect the ability of conscious individuals to make high-order discriminations. How does such an ability reflect


the efficacy of the neural states accompanying consciousexperience? Imagine an animal with primary conscious-ness in the jungle. It hears a low growling noise, and atthe same time the wind shifts and the light begins towane. It quickly runs away, to a safer location. A physi-cist might not be able to detect any necessary causal rela-tion among these events. But to an animal with primaryconsciousness, just such a set of simultaneous eventsmight have accompanied a previous experience, whichincluded the appearance of a tiger. Consciousness al-lowed integration of the present scene with the animal’spast history of conscious experience, and that integra-tion has survival value whether a tiger is present or not.An animal without primary consciousness might havemany of the individual responses that the conscious ani-mal has and might even survive. But, on average, it ismore likely to have lower chances of survival—in thesame environment it is less able than the conscious ani-mal to discriminate and plan in light of previous andpresent events.

In succeeding chapters, I will attempt to explainhow conscious scenes and qualia arise as a result of braindynamics and experience. At the outset, though, it isimportant to understand what a scientific explanationof conscious properties can and cannot do. The issueconcerns the so-called explanatory gap that arises fromthe remarkable differences between brain structure inthe material world and the properties of qualia-ladenexperience. How can the firing of neurons, however


complex, give rise to feelings, qualities, thoughts, andemotions? Some observers consider the two realms sowidely divergent as to be impossible to reconcile. Thekey task of a scientific description of consciousness is togive a causal account of the relationship between thesedomains so that properties in one domain may be un-derstood in terms of events in the other.

What such an explanation cannot and need notdo is offer an explanation that replicates or createsany particular quale or experiential state. Science doesnot do that—indeed, imagine that a gifted scientist,through an understanding of fluid dynamics and meteo-rology, came up with a powerful theory of a complexworld event like a hurricane. Implemented by a sophisti-cated computer model, this theory makes it possible tounderstand how hurricanes arise. Furthermore, with thecomputer model, the scientist could even predict mostof the occurrences and properties of individual hurri-canes. Would a person from a temperate zone withouthurricanes, on hearing and understanding this theory,then expect to experience a hurricane or even get wet?The theory allows one to understand how hurricanesarise or are entailed by certain conditions, but it cannotcreate the experience of hurricanes. In the same way, abrain-based theory of consciousness should give a causalexplanation of its properties but, having done so, itshould not be expected to generate qualia “by descrip-tion.”

To develop an adequate theory of consciousness,


one must comprehend enough of how the brain worksto understand phenomena, such as perception andmemory, that contribute to consciousness. And if thesephenomena can be causally linked, one would hope totest their postulated connections to consciousness by ex-perimental means. This means that one must find theneural correlates of consciousness. Before addressingthese issues, let us turn first to the brain.


Elements of the Brain

he human brain is the most complicated mate-rial object in the known universe. I have already saidthat certain processes within the brain provide the neces-sary mechanisms underlying consciousness. In the pastdecade or so, many of these processes have been identi-fied. Brain scientists have described an extraordinary lay-ering of brain structures at levels ranging from moleculesto neurons (the message-carrying cells of the brain), toentire regions, all affecting behavior. In describing thosefeatures of the brain necessary to our exploration I willnot go into great detail. To provide a foundation for a

14

biological theory of consciousness, however, we do needto consider certain basic information on brain structureand dynamics. This excursion will require some patienceon the reader’s part. It will be rewarded when we de-velop a picture of how the brain works.

This short survey on the brain will cover, in order, aglobal description of brain regions, some notion of theirconnectivity, the basics of the activity of neurons andtheir connections—the synapses—and a bit of thechemistry underlying neuronal activity. All this will benecessary to confront a number of critical questions andprinciples: Is the brain a computer? How is it built dur-ing development? How complex are its transactions? Arethere new principles of organization unique to the brainthat were selected during evolution? What parts of thebrain are necessary and sufficient for consciousness toemerge? In addressing these questions, I shall use the hu-man brain as my central reference. There are, of course,many similarities between our brains and those of otheranimal species, and when necessary I shall describe thesesimilarities as well as any significant differences.

The human brain weighs about three pounds. Itsmost prominent feature is the overlying wrinkled andconvoluted structure known as the cerebral cortex,which is plainly visible in pictures of the brain (Figure1). If the cerebral cortex were unfolded (making the gyri,its protrusions, and the sulci, its clefts, disappear) itwould have the size and thickness of a large table napkin.It would contain at least 30 billion neurons, or nerve

e l e m e n t s o f t h e b r a i n15

Figure 1. Relative locations of major parts of the humanbrain. The cerebral cortical mantle receives projections

from the thalamus and sends reciprocal projections back;this constitutes the thalamocortical system. Beneath themantle are three major cortical appendages—the basal

ganglia, the hippocampus, and the cerebellum.Below them is the brainstem, evolutionarily the oldest

part of the brain, which contains several diffuselyprojecting value systems.

cells, and 1 million billion connections, or synapses. Ifyou started counting these synapses right now at a rateof one per second, you would just finish counting them32 million years from now.

Neurons are connected to each other locally toform a dense network in portions of the brain calledgray matter; they communicate over longer distances viafiber tracts called white matter. The cortex itself is a six-


layered structure with different connection patterns ineach layer. The cortex is subdivided into regions thatmediate different sensory modalities, such as hearing,touch, and sight. There are other cortical regions dedi-cated to motor functions, the activity of which ulti-mately drives our muscles. Beyond the sensorimotorportions concerned with input and output, there are re-gions such as the frontal, parietal, and temporal corticesthat are connected only to other parts of the brain andnot to the outside world.

Before taking up other portions of the brain, I shallbriefly describe in simplified form the structure andfunction of neurons and synapses. Different neurons canhave a number of shapes, and there may be as many astwo hundred or more different kinds in the brain. Aneuron consists of a cell body with a diameter on theorder of thirty microns, or about one ten-thousandth ofan inch across (Figure 2). Neurons tend to be polar,with a treelike set of extensions called dendrites, and along specialized extension called an axon, which con-nects the neuron to other neurons at synapses. The syn-apse is a specialized region that links the so-called pre-synaptic neuron (the neuron that sends a signal acrossthe synapse) to a postsynaptic neuron (the neuron thatreceives the signal). The presynaptic portion of the syn-apse contains a special set of minute vesicles withinwhich are chemicals known as neurotransmitters. Neu-rons possess an electrical charge as a result of their mem-brane properties, and when a neuron is excited current


Figure 2. A diagram illustrating synaptic connectionsbetween two neurons. An action potential traveling downthe axon of the presynaptic neuron causes the release of aneurotransmitter into the synaptic cleft. The transmitter

molecules bind to receptors in the postsynaptic membrane,changing the probability that the postsynaptic neuronwill fire. (Because of the number of different shapesand kinds of neurons, this drawing is of necessity

a greatly simplified cartoon.)

flows through channels that open across the membrane.As a result, a wave of electrical potential known as anaction potential moves from the cell body down the pre-synaptic axon and causes the release of neurotransmittermolecules from vesicles into the synaptic cleft. These


molecules bind to molecular receptors or channels inthe postsynaptic cell that, acting cumulatively, can causeit to fire an action potential of its own. Thus, neuronalcommunication occurs by a combination of controlledelectrical and chemical events.

Now try to imagine the enormous numbers of neu-rons firing in various areas of the brain. Some firings arecoherent (that is, they are simultaneous), others are not.Different brain regions have different neurotransmittersand chemicals whose properties change the timing, am-plitude, and sequences of neuronal firing. To achieve andmaintain the complex patterns of dynamic activity inhealthy brains, some neurons are inhibitory, suppressingthe firing of others, which are excitatory. Most excitatoryneurons use the substance glutamate as their neurotrans-mitter, while the inhibitory neurons use GABA (gamma-aminobutyric acid). We can ignore the chemical detailsfor now and simply accept that the effects of differentchemical structures are different and that their distribu-tion and occurrence together can have significant effectson neural activity.

I started by describing the cortex. With the pictureof a polar neuron in mind, we can turn briefly to otherkey regions of the brain. One of the most importantanatomical structures for understanding the origin ofconsciousness is the thalamus. This structure, which islocated at the center of the brain, is essential for con-scious function, even though it is only somewhat largerthan the last bone in your own thumb. When nerves


from different sensory receptors serving different modal-ities (located in your eyes, ears, skin, and so on) travelto your brain, they each connect in the thalamus withspecific clusters of neurons called nuclei. Postsynapticneurons in each specific thalamic nucleus then projectaxons that travel and map to particular areas of the cor-tex. A well-studied example is the projection from theneurons of the retina through the optic nerve to the partof the thalamus called the lateral geniculate nucleus andthen to the primary visual cortical area, called V1 (for“visual area 1”).

There is one striking feature of the many connec-tions between the thalamus and the cortex: not onlydoes the cortex receive many axons from thalamic neu-rons but there are also reciprocal axonal fibers goingfrom the cortex back to the thalamus. We speak there-fore of thalamocortical projections and corticothalamicprojections. Reciprocal connections of this type aboundwithin the cortex itself; such reciprocal connections arecalled corticocortical tracts. A striking example of theseis the fiber bundle called the corpus callosum, whichconnects the two cortical hemispheres and consists ofmore than 200 million reciprocal axons. Cutting thecorpus callosum leads to a split-brain syndrome, whichin some cases can lead to the remarkable appearance oftwo separate and very different consciousnesses.

Each specific thalamic nucleus (and there aremany) does not connect directly to any of the others.Surrounding the periphery of the thalamus, however,


there is a layered structure called the reticular nucleus,which connects to the specific nuclei and which can in-hibit their activity. The reticular nucleus, it is suspected,acts to switch or “gate” the activities of the specific tha-lamic nuclei, yielding different patterns of expression ofsuch sensory modalities as sight, hearing, and touch. An-other set of thalamic nuclei called intralaminar nucleireceive connections from certain lower structures in thebrainstem that are concerned with activation of multipleneurons; these then project to many different areas ofthe cortex. The activity of these intralaminar nuclei issuspected to be essential for consciousness in that it setsappropriate thresholds or levels of cortical response—with too high a threshold, consciousness would be lost.

We may now turn to some other brain structuresthat are important to our efforts to track down the neu-ral bases of consciousness. These are large subcorticalregions that include the hippocampus, the basal ganglia,and the cerebellum. The hippocampus is an evolution-arily ancient cortical structure lined up like a pair ofcurled sausages along the inner skirt of the temporal cor-tex, one on the right side and another on the left. Incross section, each sausage looks like a sea horse, hencethe name “hippocampus.” Studies of the neural proper-ties of the hippocampus provide important examples ofsome of the synaptic mechanisms underlying memory.One such mechanism, which should not be equated withmemory itself, is the change in the strength, or efficacy,of hippocampal synapses that occurs with certain pat-


terns of neural stimulation. As a result of this change,which can be either positive for long-term potentiationor negative for long-term depression, certain neuralpathways are dynamically favored over others.

The point to be stressed is that, while synapticchange is essential for the function of memory, memoryis a system property that also depends on specific neuro-anatomical connections.

Increased synaptic strength or efficacy within apathway leads to a higher likelihood of conductionacross that pathway, whereas decreases in synapticstrength diminish that likelihood. Various patterns havebeen found for the so-called synaptic rules governingthese changes, following the initial proposals of DonaldHebb, a psychologist, and Friedrich von Hayek, aneconomist who, as a young man, thought quite a bitabout how the brain works. These scholars suggestedthat an increase in synaptic efficacy would occur whenpre- and postsynaptic neurons fired in close temporalorder. Various modifications of this fundamental rulehave been seen in different parts of the nervous system.What is particularly striking about the hippocampus,where these rules have been studied in detail, is the factthat bilateral removal of this structure leads to a loss ofepisodic memory, the memory of specific episodes orexperiences in life. A very famous patient, H. M., whosehippocampi were removed to cure epileptic seizures,could not, for example, convert his short-term memoryof events into a permanent narrative record, a condition


that was depicted dramatically in the movie Memento.It is believed that such a long-term record results whenparticular synaptic connections between the hippocam-pus and the cortex are strengthened. When these con-nections are severed, the corresponding cortical synapticchanges cannot take place and the ability to rememberepisodes over the long term is lost. Such a patient canremember episodes up to the time of the operation, butloses long-term memory thereafter. It is intriguing thatin some animals, such as rodents, hippocampal functionis necessary for memories of a sense of place. In the ab-sence of hippocampal functions, the animal cannot re-member target places that have been explored.

All of the discussion so far has focused on sensoryor cognitive functioning. The brain’s motor functions,however, are also critically important, not just for theregulation of movement, but also for forming imagesand concepts, as we shall see. A key output area is theprimary motor cortex, which sends signals downthrough the spinal cord to the muscles. There are alsomany other motor areas in the cortex, and there are nu-clei in the thalamus related to motor function as well.Another structure related to motor functions is the cere-bellum, a prominent bulb at the base of the cortex andabove the brainstem (see Figure 1). The cerebellum ap-pears to serve in the coordination and sequencing of mo-tor actions and sensorimotor loops. There is no evi-dence, however, that it participates directly in consciousactivity.


An intriguing set of structures known as the basalganglia is critically important for motor control and se-quencing. Lesions of certain structures in these nucleilead to a loss of the neurotransmitter dopamine, andthus to the symptoms of Parkinson’s syndrome. Patientswith this disease have tremors, difficulty in initiatingmotor activity, rigidity, and even certain mental symp-toms. The basal ganglia, as shown in Figure 1, are lo-cated in the center of the brain and connect to the cortexvia the thalamus. Their neural connectivity, which isradically different from that of the cortex, consists ofcircuits of successive synapses or polysynaptic loops con-necting the various ganglia. For the most part, the recip-rocal connection patterns seen in the cortex itself andbetween the cortex and the thalamus are lacking in thebasal ganglia. Moreover, most of the activity of the basalganglia is through inhibitory neurons using GABA as aneurotransmitter. Nevertheless, since inhibition of inhi-bition (or disinhibition) can occur in these loops, theycan stimulate target neurons as well as suppress theiractivity.

The basal ganglia are believed to be involved in theinitiation and control of motor patterns. It is also likelythat much of what is called procedural memory (remem-bering how to ride a bicycle, for example) and othernonconscious learned activity depends on the functionsof the basal ganglia. As we shall see later, the regulatoryfunctions of basal ganglia are also significant for formingcategories of perceptions during experience.


There is one final set of structures that is criticalin brain activity connected with learning and the main-tenance of consciousness. These are the ascending sys-tems, which my colleagues and I have called value sys-tems because their activity is related to rewards andresponses necessary for survival. They each have a differ-ent neurotransmitter, and from their nuclei of originthey send axons up and down the nervous system in adiffuse spreading pattern. These nuclei include the locuscoeruleus, a relatively small number of neurons in thebrainstem that release noradrenaline; the raphe nucleus,which releases serotonin; the various cholinergic nuclei,so-called because they release acetylcholine; the dopa-minergic nuclei, which release dopamine; and the hista-minergic system, which resides in a subcortical regioncalled the hypothalamus, a region that affects many crit-ical body functions.

The striking feature of such value systems is that,by projecting diffusely, each affects large populations ofneurons simultaneously by releasing its neurotransmit-ter in the fashion of a leaky garden hose. By doing so,these systems affect the probability that neurons in theneighborhood of value-system axons will fire after re-ceiving glutamatergic input. These systems bias neu-ronal responses affecting both learning and memory andcontrolling bodily responses necessary for survival. It isfor this reason that they are termed value systems. Inaddition, there are other loci in the brain with modula-tory functions mediated by substances called neuropep-


tides. An example is enkephalin, an endogenous opioidthat regulates responses to pain. In addition, there areother brain areas, such as the amygdala, which are in-volved in emotional responses, such as fear. For our pur-poses, these areas need not be described in detail.

To summarize our account so far, we may say that,in a gross sense, there are three main neuroanatomicalmotifs in our brains (Figure 3). The first is the thalamo-cortical motif, with tightly connected groups of neuronsconnected both locally and across distances by rich re-ciprocal connections. The second is the polysynapticloop structure of the inhibitory circuits of the basal

Figure 3. Fundamental arrangements of three kinds ofneuroanatomical systems in the brain. The top diagramshows the gross topology of the thalamocortical system,

which is a dense meshwork of reentrant connectivitybetween the cortex and the thalamus and among different

cortical areas. The middle diagram shows the longpolysynaptic loops connecting the cortex with subcortical

structures such as the basal ganglia. In this case, theseloops go from the basal ganglia to the thalamus, thence tothe cortex and back from the target areas of cortex to theganglia. These loops are, in general, not reentrant. Thebottom diagram shows one of the diffusely projecting

value systems, in which the locus coeruleus distributes a“hairnet” of fibers all over the brain. These fibers release

the neuromodulator noradrenaline when the locuscoeruleus is activated.


ganglia. The third consists of the diffuse ascending pro-jections of the different value systems. Of course, thisgeneralization is a gross oversimplification, given the ex-quisite detail and individuation of neural circuitry. Butas we shall see, it provides a useful simplification; wecan dispose of it once we have seen its uses.

So much for simplicity. The picture I have paintedso far only hints at the remarkably complex dynamicsof the neural structures of the brain. After staring at thegross layout of brain regions in Figure 1 and under-standing the synapse pictured in Figure 2, close youreyes and imagine myriad neural firings in millions ofpathways. Some of this neural activity would occur atcertain frequencies while others would show variable fre-quencies. Bodily activity and signals from the environ-ment and the brain itself would modify which of thepathways were favored over others as a result of changesin synaptic strength. Although I hardly expect the readerto be able to visualize precisely the hyperastronomicalnumbers of neural patterns in detail, perhaps this exer-cise will yield a further appreciation of the brain’s com-plexity.

We are now in a position to address some of thequestions posed at the beginning of this chapter. Con-sider the question of whether the brain is a computer. Ifwe examine how neural circuits are built during animaldevelopment, this would seem unlikely. The brain arisesduring development from a region of the embryo calledthe neural tube. Progenitor cells (cells that are precursors


to neurons and support cells called glia) move in certainpatterns to make various layers and patterns. As theydifferentiate into neurons, many also die. From the verybeginning of neuroanatomy, there are rich statisticalvariations in both cell movement and cell death. As aresult, no two individuals, not even identical twins, pos-sess the same anatomical patterns.

In the earliest stages of development, the cellularorganization characteristic of a species is controlled byfamilies of genes among which are the so-called Hoxgenes and Pax genes. But at a certain point, the controlof neural connectivity and fate becomes epigenetic; thatis, it is not prespecified as “hardwiring,” but rather isguided by patterns of neural activity. Neurons that firetogether wire together. While, at earlier stages, patternedcell movement and programmed cell death determineanatomical structure, the movement and death of indi-vidual neurons are nonetheless statistically variable orstochastic. The same holds for which particular neuronsconnect to each other at later stages. The result is a pat-tern of constancy and variation leading to highly indi-vidual networks in each animal. This is no way to builda computer, which must execute input algorithms oreffective procedures according to a precise prearrangedprogram and with no error in wiring.

There are other, even more trenchant reasons forrejecting the idea of digital computation as a basis forbrain action. As we shall see later, what would be lethalnoise for a computer is in fact critical for the operation


of higher-order brain functions. For the moment,though, let us consider some other aspects of brain com-plexity and its relation to brain structure and function.

A review of what I have said about the overall ar-rangement of brain areas might tempt one to concludethat the key to brain function is modularity. Since thereare regions that are functionally segregated for vision(even for color, movement, and orientation), for exam-ple, or similarly for hearing or touch, we might betempted to conclude that specific brain action is mainlythe result of the specialized functioning of these isolatedlocal parts or modules. If pushed to higher levels, thissimple notion results in phrenology, the picture of local-ized separate brain faculties first proposed by Franz Jo-seph Gall. We now know that modularity of this kindis indefensible. The alternative picture, that the brainoperates only as a whole (the holistic view), will also notstand up to scrutiny.

The notion of modularity is based on an overlysimple interpretation of the effects of ablation of partsof the brain, either by animal experiments or as a resultof a stroke, or of surgery for epilepsy. It is clear, forexample, that ablation of cortical area V1 leads to blind-ness. It does not follow, however, that all the propertiesof vision are assured by the functioning of V1, whichis the first cortical area in a series making up the visualpathway. Similarly, although modern imaging tech-niques reveal certain areas of the brain that are active incertain tasks, it does not follow that the activity of such


areas is the sole cause of particular behaviors. Necessityis not sufficiency. But the contrary or holistic argumentis not tenable either—one must account for both inte-gration and differentiation of brain activity. This willbe one of our main tasks in proposing a global braintheory. As we shall see later, the long-standing argumentbetween localizationists and holists dissolves if one con-siders how the functionally segregated regions of thebrain are connected as a complex system in an intricatebut integrated fashion. This integration is essential tothe emergence of consciousness.

This reasoning is critical to understanding the rela-tionship between brain function and consciousness. Ofcourse, there are areas of the brain that if damaged orremoved will lead to permanent unconsciousness. Onesuch area is the midbrain reticular formation. Anotheris the region of the thalamus containing the intralaminarnuclei. These structures are not the site of consciousness,however. As a process, consciousness needs their activ-ity, but to account for the Jamesian properties of con-sciousness requires a much more dynamic picture in-volving integration of the activities of multiple brainregions. We are now in a position to lay the groundworkfor just such a picture by considering a global brain the-ory that accounts for the evolution, development, andfunction of this most complex of organs.


Neural DarwinismA G L O B A L B R A I N T H E O R Y

here is one simple principle that governs howthe brain works: it evolved; that is, it was not designed.As stated, this principle sounds almost simple-minded,doesn’t it? But we must not forget that, although evolu-tion is not intelligent, it is enormously powerful. Thepower comes from natural selection acting in complexenvironments over eons of time. A key idea developedby Darwin is embedded in his notion of populationthinking: functioning structures and whole organismsemerge as a result of selection among the diverse variant

32

individuals in a population, which compete with one an-other for survival. I hold this notion to be central, notonly in considering how the brain has evolved, but alsoin thinking about how it develops and functions.Applying population thinking to understanding how thebrain works leads to a global theory, called neural Dar-winism or the theory of neuronal group selection.

What do we mean by the term “global” and whydo we need a global brain theory? An explanation ofconsciousness will necessarily require an understandingof perception, memory, action, and intention—inshort, an overall understanding of how the brain worksthat goes beyond the functioning of one brain regionor another. Given the richness, variety, and range ofconscious experience, it is also important to constructa brain theory that is principled and compatible withevolution and development. By principled, I mean atheory that describes the principles governing the majormechanisms by which the brain deals with informationand novelty. One such theory or model is the idea thatthe brain is like a computer or Turing machine. In con-trast to such an instructive model, which relies on pro-grams and algorithms, models based on populationthinking rely on selection of particular elements or statesfrom a large repertoire of variant elements or states. Ex-planations of consciousness based on one or the otherof these two kinds of models differ greatly. By now, itshould be no mystery that I prefer selectional modelsbased on population thinking.

n e u r a l d a r w i n i s m33

The reason population thinking is important in de-termining how the brain works has to do with the ex-traordinary amount of variation in each individualbrain. This is true at all levels of structure and function.Different individuals have different genetic influences,different epigenetic sequences, different bodily re-sponses, and different histories in varying environments.The result is enormous variation at the levels of neuronalchemistry, network structure, synaptic strengths, tem-poral properties, memories, and motivational patternsgoverned by value systems. In the end, there are obviousdifferences from person to person in the contents andstyles of their streams of consciousness. The variabilityof individual nervous systems was commented on by thedistinguished neuroscientist Karl Lashley, who admittedthat he had no ready explanation for the existence of somuch variation. Even though there are general patternsexhibited by the brain in the face of this variation, itcannot be dismissed as mere noise. There is too muchof it, and it exists at too many levels of organization—molecules, cells, and circuits. It is simply not likely thatevolution, like a computer programmer dealing withnoise, could have devised multiple error-correctingcodes to assure preservation of patterns in the brain bycounteracting this enormous variation.

An alternative way of confronting neural variabilityis to consider it fundamental and to assume that theindividual local differences within each brain make uppopulations of variants. In this case, selection from such


a population of variants could lead to patterns even un-der unpredictable circumstances, provided that someconstraint of value or fitness was satisfied. In evolution,fitter individuals survive and have more progeny. In theindividual brain, those synaptic populations that matchvalue systems or rewards are more likely to survive orcontribute more to the production of future behavior.

This view is in sharp contrast to computer modelsof the brain and mind. According to these models, sig-nals from the environment carry input information thatis unambiguous, once contaminating noise is averagedaway or otherwise dealt with. These models assume thatthe brain has a set of programs, or so-called effective pro-cedures, which are capable of changing states based onthe information carried by the inputs, yielding function-ally appropriate outputs. Such models are instructive inthe sense that information from the world is assumed toelicit the formation of appropriate responses based onlogical deduction. These models do not deal, however,with the fact that inputs to the brain are not unambigu-ous—the world is not like a piece of tape with a fixedsequence of symbols for the brain to read. I have alreadymentioned the challenge to computer models of thebrain posed by the richly variable circuitry of real brains.

There is also a set of functional issues that makecomputer models unlikely. For example, the mappedconnections from the sense of touch in the handthrough the thalamus to the region of somatosensorycortex are variable and plastic, even in adults. The sub-


regions in the somatosensory cortex mapping the fingersdynamically shift all their boundaries as a result of exces-sive use of even one finger—a shift in the context of use.Similar phenomena reflecting such context dependenceand dynamic circuit variation are seen for other senses.Furthermore, in sensory systems such as that for vision,there are multiple cortical regions that are each func-tionally segregated, for example, for color, movement,orientation, and so on. These functionally specializedareas can exceed thirty in number and are distributedall over the brain. Yet there is no superordinate area orexecutive program binding the color, edge, form, andmovement of an object into a coherent percept. Thisbinding is not explicable by invoking a visual computerprogram operating according to the principles of artifi-cial intelligence. A coherent percept in fact neverthelessemerges in various contexts, and explaining how this oc-curs constitutes the so-called binding problem. A globalbrain theory must provide a cogent solution to thisproblem by proposing an appropriate mechanism. Itwill soon become clear that such a solution is central toour understanding of consciousness.

To emphasize the dependence of perception oncontext, we may call upon the huge phenomenologyof illusions, visual and otherwise. One example is theKanizsa pattern, which consists of the angular portionsof a triangle, disconnected, but appears to show an over-lying triangle with sharp boundaries (Figure 4). Yetthere is no true energy difference in the light that is


Figure 4. Illusory contours in a Kanizsa triangle. Mostpeople report the appearance of a distinct triangular shapeand an increase in apparent luminance within the triangle,but neither of these features exists in the physical image.

received from the two sides of the contour that is per-ceived. Such a contour is called “illusory.” The brainconstructs the contour, which, by the way, is not neces-sarily a straight line but can be curved depending on thecontext of the particular figure used.

Many other functional responses of the perceivinganimal could be described to illustrate why an a prioriprogram is not a likely explanation for physiological orpsychological properties. I shall mention only two more.The first is the remarkable tendency of brains to seekout closure and avoid gaps. In daily life, for example,you do not see the blind spot in your visual field occa-sioned by the presence of the optic nerve near the centerof your retina. Even more striking phenomena come


from the field of neuropsychology, which, among otherthings, studies responses to strokes. This field is repletewith examples of closure phenomena that can even bedelusional. A most exotic example is anosognosia, a syn-drome in which a paralyzed patient does not recognizethe existence of paralysis even if it involves his or herentire left side. In such cases, we see extraordinary adap-tation and integration by the damaged brain as it re-sponds to the loss of cortical areas.

In addition to construction and closure, and possi-bly in connection with them, the brain’s capacity to gen-eralize is astonishing. A case in point is the ability ofpigeons, when appropriately rewarded, to look at nu-merous photographs of various fish species in differentscales and contexts and learn to positively recognize thesimilarity in the photographs. Pigeons trained at thistask can recognize that these diverse pictures have some-thing in common more than 80 percent of the time. Itis highly unlikely that this behavior is the result of afixed template or a set of predetermined algorithms inthe brains of pigeons. Nor can it be explained by naturalselection for the positive recognition of fish. Pigeonsneither evolve with fish nor live with them, and theydon’t eat them either.

I could cite many more examples ranging from thedevelopmental anatomy of the brain to the individualvariation of brain scans in humans carrying out similartasks. But the conclusion is clear: the brains of higher-level animals autonomously construct patterned re-


sponses to environments that are full of novelty. Theydo not do this the way a computer does—using formalrules governed by explicit, unambiguous instructions orinput signals. Once more, with feeling: the brain is nota computer, and the world is not a piece of tape.

If the brain is in fact not a computer and the worldis not a piece of tape, how can the brain operate so as toyield adaptive and patterned responses? As I have alreadysuggested, the answer lies in a selectionist theory that Ihave called the theory of neuronal group selection, orTNGS (Figure 5). This theory has three tenets: (1) De-velopmental selection—during the early establishmentof neuroanatomy, epigenetic variations in the patternsof connections among growing neurons create reper-toires in each brain area consisting of millions of variantcircuits or neuronal groups. The variations arise at thelevel of synapses as a result of the fact that neurons thatfire together wire together during the embryonic andfetal stages of development. (2) Experiential selection—overlapping this first phase of selection and after the ma-jor neuroanatomy is built, large variations in synapticstrengths, positive and negative, result from variationsin environmental input during behavior. These synapticmodifications are subject to the constraints of value sys-tems described in the previous chapter. (3) Reentry—during development, large numbers of reciprocal con-nections are established both locally and over longdistances. This provides a basis for signaling betweenmapped areas across such reciprocal fibers. Reentry is


Figure 5. The three main tenets of the theory of neuronalgroup selection, or neural Darwinism: (1) Developmentalselection leads to a highly diverse set of circuits, one of

which is shown. (2) Experiential selection leads to changes inthe connection strengths of synapses, favoring some pathways(thickened black lines) and weakening others (dashed lines).

(3) Reentrant mapping, in which brain maps are coordinatedin space and time through ongoing reentrant signaling acrossreciprocal connections. The black dots in the maps on theright indicate strengthened synapses. As a result of (1) and

(2), a myriad of circuits and functioning pathways is createdconstituting a repertoire for selectional events. The further

and ongoing events of reentry in (3) must be thought of asdynamic and recursive, mapping the maps over time.

the ongoing recursive interchange of parallel signalsamong brain areas, which serves to coordinate the activi-ties of different brain areas in space and time. Unlikefeedback, reentry is not a sequential transmission of anerror signal in a simple loop. Instead, it simultaneouslyinvolves many parallel reciprocal paths and has no pre-scribed error function attached to it.

The consequence of this dynamic process is thewidespread synchronization of the activity of widely dis-tributed neuronal groups. It binds their functionally seg-regated activities into circuits capable of coherent out-put. In the absence of logic (the organizing principle ofcomputers as instructive systems), reentry is the centralorganizing principle that governs the spatiotemporal co-ordination among multiple selectional networks of thebrain. This solves the binding problem that I mentionedearlier. Through reentry, for example, the color, orienta-tion, and movement of a visual object can be integrated.No superordinate map is necessary to coordinate andbind the activities of the various individual maps thatare functionally segregated for each of these attributes.Instead, they coordinate by communicating directlywith each other, through reentry.

The three tenets of the TNGS together form a se-lectional system. Prominent examples of selectional sys-tems include evolution, the immune system, and com-plex nervous systems. All follow a set of three guidingprinciples. The first principle assumes a means for gen-erating diversity in a population of elements, whether


of individuals or of cells. The second is a means allowingextensive encounters between individuals in a variantpopulation or repertoire and the system that is to berecognized, whether it is an ecological environment, aforeign molecule, or a set of sensory signals. The thirdprinciple is some means to differentially amplify thenumber, survival, or influence of those elements in thediverse repertoire that happen to meet selective criteria.In evolution, these are criteria of fitness allowing thedifferential survival and breeding of certain individu-als—the process of natural selection itself. In immunity,amplification occurs through the enhanced division ofjust those clones of immune cells having antibodies ontheir surface that bind particular foreign molecules orantigens well enough to exceed a certain critical energyof binding. In neural systems, amplification consists ofenhancing the strengths of those synapses and circuitsof neuronal groups that meet the criteria set by valuesystems. It is the neuronal groups made up of excitatoryand inhibitory neurons in particular anatomical patternsrather than individual neurons that are selected.

Notice that while these three different selectionalsystems obey similar principles, they use different mecha-nisms to achieve successful matching to various unfore-seen inputs. Evolution is, of course, special and over-arching because it is also responsible for actuallyselecting the different mechanisms used by the immuneand nervous systems. It tends to favor those individualsthat successfully utilize such mechanisms to improve


their fitness and allow more of their progeny to sur-vive.

Since the proposal of the TNGS in 1978, a grow-ing body of evidence has supported the notion that neu-ronal groups connected by reentrant interactions are theselectional units in higher-level brains. This evidence ispresented in a number of books and papers and willnot be reviewed here. Instead, I will consider certainconsequences of the theory that are particularly impor-tant for understanding the mechanisms underlying con-sciousness.

One important consequence is that the brain is soversatile in its responses because those responses are de-generate. Degeneracy is the ability of structurally differ-ent elements of a system to perform the same functionor yield the same output. A clear-cut example is seen inthe genetic code. The code is made up of triplets ofnucleotide bases, of which there are four kinds: G, C,A, and T. Each triplet, or codon, specifies one of thetwenty different amino acids that make up a protein.Since there are sixty-four different possible codons—actually sixty-one, if we leave out three stop codons—which makes a total of more than one per amino acid,the code words are degenerate. For example, the thirdposition of many triplet codons can contain any one ofthe four letters or bases without changing their codingspecificity. If it takes a sequence of three hundred co-dons to specify a sequence of one hundred amino acidsin a protein, then a large number of different base se-


quences in messages (approximately 3100) can specify thesame amino-acid sequence. Despite their different struc-tures at the level of nucleotides, these degenerate mes-sages yield the same protein.

Degeneracy is a ubiquitous biological property. Itrequires a certain degree of complexity, not only at thegenetic level as I have illustrated above, but also at cellu-lar, organismal, and population levels. Indeed, degener-acy is necessary for natural selection to operate and it isa central feature of immune responses. Even identicaltwins who have similar immune responses to a foreignagent, for example, do not generally use identical combi-nations of antibodies to react to that agent. This is be-cause there are many structurally different antibodieswith similar specificities that can be selected in the im-mune response to a given foreign molecule.

Degeneracy is particularly important in helping tosolve major problems in complex nervous systems. Ihave already mentioned the binding problem. How canit be that, despite the absence of a computer program,executive function, or superordinate map, up to thirty-three functionally segregated and widely distributed vi-sual maps in the brain can nevertheless yield perceptionthat coherently binds edges, orientations, colors, andmovement into one perceptual image? How do differentmaps for color, orientation, object movement, and soon correlate or coordinate their responses? As I suggestedabove, the answer lies in mutual reentrant interactionsthat, for a time, link various neuronal groups in each


map to those of others to form a functioning circuit.Simulations show that the neurons that yield such cir-cuits fire more or less in phase with each other, or syn-chronously. But in the next time period, different neu-rons and neuronal groups may form a structurallydifferent circuit, which nevertheless has the same out-put. And again, in the succeeding time period, a newcircuit is formed using some of the same neurons, aswell as completely new ones in different groups. Thesedifferent circuits are degenerate—they are different instructure but they yield similar outputs to solve thebinding problem (Figure 6).

Within each particular circuit, the different neu-ronal groups fire synchronously. The different circuitsyielding the same output are not, however, synchronousor in phase with each other, nor do they have to be.As a result of reentry, the properties of synchrony andcoherency allow more than one structure to give a simi-lar output. As long as such degenerate operations occurin succession to link distributed populations of neuronalgroups, there is no need for an executive or superordi-nate program as there would be in a computer.

The formulation of a global brain theory like theTNGS, while essential to understanding how the brainworks, does not solve all of the detailed mechanisticproblems related to the local operations of networks inthe various nuclei and regions of the brain. But it doesremove the paradoxes that arise if one assumes that thebrain functions like a computer. One such paradox


Figure 6. Illustration of the degeneracy of reentrant circuitsin the brain. Even though the three overlapping circuits inA, B, and C are different, as shown by the shading, they

can yield a similar output over some period of time.

would have us imagine a cell with a designated categori-cal function that dominates the function of all subordi-nate neurons connected to it—for example, a cell thatfires when you think of a particular person, a so-calledgrandmother cell. Such a cell is not necessary in thistheory. Different cells can carry out the same functionand the same cell can, at two different times, carry outdifferent functions in different neuronal groups. More-over, given the selectional nature of higher-order inter-actions in the brain, one does not have to invoke a ho-


munculus, a little man who lives in the brain, tointerpret the meaning of a percept. Just as Darwin’s the-ory of natural selection disposed of the argument fromdesign, the TNGS disposes of the need for either a fixedinstructional plan or a homunculus in the head.

These issues are directly relevant to my next task,which is to show how the principles and mechanismsof the TNGS can be used to understand the origin ofconsciousness.


The Mechanisms ofConsciousness

y fundamental assumption has been thatthe process of consciousness arises from the workings ofthe brain. I must show how that can occur as an evolu-tionary event connecting previously evolved capabilitieswith new structural and functional features that emergeas a result of natural selection. To do so, I must dissectthe necessary components whose interactions result inthe appearance of primary consciousness—the ability toconstruct a scene in a discriminative fashion. Thus, be-fore proposing mechanisms of consciousness, I will con-

48

sider the contributory brain processes that are essentialfor these mechanisms to operate.

One of the most basic processes in higher brainsis the ability to carry out perceptual categorization—to“make sense” of the world. This ability allows an ani-mal to carve up the world of signals coming fromthe body and the environment into sequences thatresult in adaptive behavior. For example, we continuallytake in the parallel and multiple visual signals from aroom and categorize them as coherent stable objects(“chairs,” “tables,” and so forth). A cat might do sucha categorization but with different perceptual andmotor responses (it might jump on the object we calla table). And a cockroach might treat that sameobject as a place to hide in the darkness of the table’sunderside.

In the mammalian nervous system, perceptual cat-egorization is carried out by interactions between sen-sory and motor systems in what I have called globalmappings. A global mapping is a dynamic structure con-taining various sensory maps, each with different func-tionally segregated properties, linked by reentry. Theseare linked in turn by non-reentrant connections to mo-tor maps and subcortical systems such as the cerebellumand basal ganglia. The function of a global mappingis first to sample the world of signals by movementand attention and then to categorize these signals as co-herent through reentry and synchronization of neuronalgroups. Such a structure, consisting of both sensory and

t h e m e c h a n i s m s o f c o n s c i o u s n e s s49

motor components, is the main basis for perceptual cat-egorization in higher brains.

While perceptual categorization is fundamental, itcannot itself give rise to generalization across various sig-nal complexes to yield the properties that the signalshave in common. For such generalization, the brainmust map its own activities, as represented by severalglobal mappings, to create a concept—that is, to makemaps of its perceptual maps. For example, to registerforward motion, a cat’s nervous system might map itsown activities as “cerebellum and basal ganglia active inpattern a, premotor and motor regions active in patternb, and submodalities in visual maps active in patternsx, y, and z.” Note that although I have, for purposes ofillustration, expressed this generalized mapping in theform of a proposition (or verbal expression), the opera-tion in a cat’s brain is obviously non-propositional.Higher-order cortical maps in the prefrontal, parietal,and temporal areas are likely to carry out this construc-tion, which might correspond to a “universal,” a conceptof forward motion. No linear sum of global mappingscould give rise to such a generalization. Instead, general-ization arises by abstracting certain features of such map-pings by means of higher-order maps.

Perceptual categorization and concept formationwould not be adaptive to an animal in the absence ofmemory, and, as we shall see, the understanding ofmemory is essential for formulating a theory of con-sciousness. According to the TNGS, memory is the ca-


pacity to repeat or suppress a specific mental or physicalact. It arises as a result of changes in synaptic efficacy(or synaptic strength) in circuits of neuronal groups.After such changes have occurred, they tend to favorthe recruitment of certain of these circuits to yield re-enactment. This recruitment may occur in more thanone way—that is, in a degenerate fashion. As we shallsee, certain forms of memory require either relativelyfast changes in synaptic efficacy or at least the ongoingactivity of particular neural circuits in time periods ofless than about one-third of a second. Other forms ofmemory require slower, but more stable, changes insynaptic strength.

Students of memory have categorized variousmemory systems in a useful fashion. They distinguishlong-term memory from short-term, or working, mem-ory by taking into account how long a particular mem-ory lasts and the structures it depends on. Brain scien-tists also distinguish between procedural memory—thatwhich reflects motor learning and its complex acts—from episodic memory, the ability to carry out long-term recall of sequences of events or narratives. As I havealready mentioned, episodic memory depends on inter-actions between the hippocampus and the cerebral cor-tex. While these various classifications are very impor-tant and useful, it is likely that there are many additionalsystems of memory that remain to be described. More-over, much remains to be done to uncover the interac-tions among the various memory systems.


There are a number of additional issues that wemust clarify to understand how memory operates inhigher brains. For example, memory cannot simply beequated with synaptic change, although changes in syn-aptic strength are essential for it. Instead, memory is asystem property reflecting the effects of context and theassociations of the various degenerate circuits capable ofyielding a similar output. Thus, each event of memoryis dynamic and context-sensitive—it yields a repetitionof a mental or physical act that is similar but not identi-cal to previous acts. It is recategorical: it does not repli-cate an original experience exactly. There is no reasonto assume that such a memory is representational in thesense that it stores a static registered code for some act.Instead, it is more fruitfully looked on as a property ofdegenerate nonlinear interactions in a multidimensionalnetwork of neuronal groups. Such interactions allow anon-identical “reliving” of a set of prior acts and events,yet there is often the illusion that one is recalling anevent exactly as it happened.

Two analogies are useful in clarifying this point.A representational memory would be like a coded in-scription cut into a rock that is subsequently broughtback into view and interpreted. A nonrepresentationalmemory would be like changes in a glacier influencedby changes in the weather, which are interpreted as sig-nals. In the analogy, the melting and refreezing of theglacier represent changes in the synaptic response, theensuing different rivulets descending the mountain ter-


rain represent the neural pathways, and the pond intowhich they feed represents the output. Successive melt-ings and refreezings due to changes in the weather couldlead to a degenerate set of paths of water descending inthe rivulets, some of which might join and associate innovel ways. Occasionally, an entirely new pond mightbe created. In no case, however, is it likely that the samedynamic pattern will be repeated exactly, althoughthe general consequences of changes in the pondbelow—the output state—could be quite similar. Inthis view, memories are necessarily associative and arenever identical. Nevertheless, under various constraintsthey can be sufficiently effective in yielding the sameoutput.

Recognizing that a dynamic memory system oper-ates within the brain’s selectional framework impliesthat it will be influenced by the changes in neural inputsthat come from that brain’s value systems. Indeed, no-tice that the mechanisms leading to perceptual categori-zation—global mappings, concept formation, and dy-namic short-term memory—all call upon interactions ofthe three major motifs of global neural systems that werediscussed in reviewing the neuroanatomy of higherbrains in Chapter 3. These are the thalamocorticalmaps, the subcortical organs concerned with temporalsuccession (the hippocampus, basal ganglia, and cerebel-lum), and the diffuse ascending value systems. To reflectthese interactions, I have called the central memory sys-tem a value-category memory system, one in which the


constraints of value systems can determine the degreeand extent of recall and output. Animals without con-sciousness still utilize all of the above systems, but theywould lack the critical interactions leading to conscious-ness. In fact, it is a central tenet of the extended TNGS(the theory applied to consciousness) that the develop-ment of all these systems was a necessary evolutionaryprecursor of conscious activity.

We may now ask the critical question: what is thesufficient evolutionary event leading to the emergence ofconsciousness? The thesis I am proposing is that, at apoint in evolutionary time corresponding to the transi-tion between reptiles and birds and reptiles and mam-mals, a new reciprocal connectivity appeared in thethalamocortical system. Massively reentrant connectiv-ity developed between the cortical areas carrying outperceptual categorization and the more frontal areasresponsible for value-category memory based on fastchanges in synaptic strength. Cortical reentry was medi-ated by the emergence of several grand systems of reen-trant corticocortical connections linking distributed ar-eas of the cortex. At the same time there was an increasein the reentrant connectivity with the thalamus, as wellas an increase in the number of thalamic nuclei. Thereentrant connections between the thalamus and cortexwere enhanced, both for the specific thalamic nuclei andthe intralaminar nuclei described in Chapter 3, whilethe reticular nucleus of the thalamus developed en-hanced inhibitory circuits by which it connected to the


specific nuclei. This allowed the activity of the reticularnucleus to gate or select various combinations of theactivity of those specific thalamic nuclei correspondingto different sensory modalities. The intralaminar nuclei,which send diffuse connections to most areas of the cor-tex, helped to synchronize the new thalamocortical re-sponses and regulate the overall levels of activity in thesemultiple reentrant systems (Figure 7).

These dynamic reentrant interactions in the thala-mocortical system must be thought of as successive intime—new perceptual categorizations are reentrantlyconnected to memory systems before they themselvesbecome part of an altered memory system. This boot-strapping between memory and perception is assumedto be stabilized within time periods ranging from hun-dreds of milliseconds to seconds—the so-called speciouspresent of William James. I have called this period “theremembered present” to point up the dynamic interac-tion between memory and ongoing perception that givesrise to consciousness.

What is the consequence of this evolutionary de-velopment in which value-category memory was dy-namically linked to perceptual categorization? It is theability to construct a complex scene and to make dis-criminations between components of that scene. As ananimal moves, engaging many global mappings in re-sponse to the world around it, the ongoing parallel sig-nals reentrantly connecting different sensory modalitieslead to correlations among complexes of perceptual cate-


Figure 7. Reentrant pathways leading to primaryconsciousness. Two main kinds of signals are critical—

those from “self,” constituting value systems and regulatoryelements of the brain and body along with their sensorycomponents, and those from “nonself,” signals from the

world that are transformed through global mappings. Signalsrelated to value and categorized signals from the outside

world are correlated and lead to memory, which is capable ofconceptual categorization. This “value-category” memory islinked by reentrant paths (the heavy lines) to the currentperceptual categorization of world signals. This reentrant

linkage is the critical evolutionary development that resultsin primary consciousness. When it occurs across many

modalities (sight, touch, and so forth), primary consciousnessis of a “scene” made up of responses to objects and events,

some of which are not necessarily causally connected toeach other. An animal with primary consciousness cannonetheless discriminate and connect these objects andevents through the memory of its previous value-laden

experience. This ability enhances its survival value.

gories stimulated by objects and events. The ability tocreate a scene by such reentrant correlations betweenvalue-category memory—reflecting earlier categoriza-tions—and similar or different perceptual categories isthe basis for the emergence of primary consciousness.

Some of the earliest categorizations are related tosignals from the animal’s own body and brain. Thesesignals come from autonomic and homeostatic systemsthat regulate vital organs and interactions of physiologi-cal functions such as breathing, eating, and hormonalchanges. They are called autonomic because they do notdepend on conscious control and are homeostatic be-cause they compensate for changes in a balanced fash-ion. Other bodily signals come from muscles and jointsand systems related to balance—so-called kinestheticand proprioceptive systems. All of these systems con-tinue to operate in the life of the animal, providing acentral referential set of signals and perceptual categoriesto that individual. Signals from such “self” systems be-gin even before birth and remain as a central feature ofprimary consciousness. The salience of various elementscontributing to the scene is governed by memories con-ditioned by the history of reward and punishment dur-ing the animal’s past behavior. Such a history plays a keyrole in emotional responses and their associated feelings.

The ability to construct a conscious scene in a frac-tion of a second is the ability to construct a rememberedpresent. Note that the causal or physical connection be-tween several incoming signals is not necessarily a decid-


ing issue in the animal’s response to this construction.For example, as I have already mentioned, an animal inthe jungle sensing a shift in sounds while the light isdiminished in its immediate surroundings may flee,even if there is no causal correlation between these twoinputs. It is sufficient that the combination of such si-multaneous inputs in the past value-dependent historyof that animal was previously accompanied by the pres-ence, say, of a tiger. An animal without the gift of pri-mary consciousness might survive for some time in sucha niche but could not make the same discriminations onthe basis of its rapidly changing value-category memory.Eventually, that animal is less likely to survive. By con-trast, an animal with the ability to construct a scene canhave a greater discriminatory capacity and selectivity inchoosing its responses to novel and complex environ-ments. The efficacity of its conscious systems and theirpossible contribution to increased fitness rests in theirenormously increased discriminatory capacity.

The brief account given here refers to the emer-gence of mechanisms for primary consciousness. Theyare consistent with the observation that consciousness isan active process. As I shall discuss later, the subsequentevolution of additional reentrant circuits permitting theacquisition of semantic capability, and finally language,gave rise to higher-order consciousness in certain higherprimates, including our hominine ancestors (and argu-ably a number of other ape species). Higher-order con-sciousness confers the ability to imagine the future, ex-


plicitly recall the past, and to be conscious of being con-scious. While I shall not consider the details at thispoint, it is necessary to use the example of higher-orderconsciousness from time to time to discuss matters sig-nificant for the understanding of primary consciousness.This is the case because the presence of higher-orderconsciousness allows direct report to an experimenter.As a result, he or she can probe conscious states andtheir neural correlates with greater assurance. Animalsother than humans cannot report their conscious statesin the absence of language. Nevertheless, there is amplereason to believe that, on the basis of their behavior andthe homologous structures and similar functions of theirnervous systems, other animals experience primary con-sciousness. So the order of study and discussion must,to some extent, proceed from humans “downward.” Itmust never be forgotten, however, that primary con-sciousness is the fundamental state, for without it, therecould be no higher-order consciousness.


Wider Than the SkyQ U A L I A , U N I T Y , A N D C O M P L E X I T Y

y emphasizing the neuroan-atomical and dynamic properties of the brain in seekingthe mechanisms of consciousness, I may appear to besidestepping some fundamental issues related to con-scious experience. How, for example, does our neuralmodel fit with the experienced properties of a conscioussubject? I believe that the issue is best clarified by stress-ing the neural mechanisms first, and then going backand forth between phenomenal issues and these mecha-nisms to show their consistency with each other.

60

One extraordinary phenomenal feature of con-scious experience is that normally it is all of a piece—it is unitary. Any experienced conscious moment simul-taneously includes sensory input, consequences of mo-tor activity, imagery, emotions, fleeting memories,bodily sensations, and a peripheral fringe. In any ordi-nary circumstances it does not consist of “just this pen-cil with which I am writing,” nor can I reduce it tothat. Yet, at the same time, one unitary scene flows andtransforms itself into another complex but also unitaryscene. Alternatively, it can shift into diffuse reverie orinto high focal attention by choice or under stress.

One way of describing this is to say that while con-scious experience is highly integrated, it is at the sametime highly differentiated. In short time periods, it canrange phenomenally over a multitude of inner states.This apparently unending change and changeabilitynonetheless cannot at any one time be dissected intounique isolated parts by the person experiencing thesesubjective states. This is not to deny that consciousnesscan be modulated by focal attention. We will discusssuch focused narrowing of the conscious scene laterwhen we consider the relation between conscious andnonconscious activities.

The subjective experience of rich inner consciousstates must be contrasted with the inability of a con-scious subject to carry out three or more conscious actssimultaneously—for example, type text, recite a poem,and answer a quiz, all at the same time. This inability

w i d e r t h a n t h e s k y61

to execute multiple tasks simultaneously has causedsome to consider consciousness to be of very limitedutility. But in fact, it is likely that this apparent limita-tion derives from the evolutionary necessity that motoractions and plans not be interrupted before completion.Moreover, the view that “chunking” simultaneous con-scious acts at best into only two or three units revealsa limit to the efficacy of the conscious state misconstruesthe relationship of that state to instrumental acts in thefuture. As we shall see, a major function of conscious-ness and its underlying neural mechanisms is planningand rehearsal and, for these, the multifarious complexityof successive inner states is just what is required. Forplanning, we must rehearse distinctions that make a dif-ference from an individual’s vantage point; that is, fromthe first-person view of a subject. Carrying out motoracts or other performances often requires conscious re-hearsal, but, after learning, such acts are more effectivelyexecuted by the subject without direct conscious super-vision, except when novel circumstances arise. It is notsurprising that an attempt to execute two or more ofany such acts that require completion is likely to be in-terrupted by conscious intervention.

How about phenomenal experience itself ? What isit that appears to the conscious subject? What does heor she feel? The term “quale” has been applied to theexperiencing of feeling—say, of green, or warmth, orpain. Philosophers have considered the understandingof qualia to be a critical problem in consciousness re-


search. Some of their concerns relate to the apparentdiscrepancy in kind between neural activity and thestructure and “feel” of qualia. I shall devote some spaceto this issue, which, simply stated, explores what it islike to be a conscious individual in a particular species—or, as the philosopher Thomas Nagel has put it, “Whatis it like to be a bat?”

To get at this issue, a number of subsidiary issuesmust be addressed. The first relates to the notion thatneural activity, as measured and understood by a scien-tific observer, has none of the properties we ascribe toqualia. Here it is useful to remember that the consciousexperience of qualia is a process. The dynamic structuralorigin of properties, even conscious properties, need notresemble the properties it gives rise to: an explosion doesnot resemble an explosive. A second issue concerns sub-jectivity and the first-person perspective. Consciousnessis a process that is tied to an individual body and brainand to their history. From an observational point ofview, the first-person experience is not written in trans-ferable currency that is completely negotiable by a third-person scientific observer. But it is a reasonable startingpoint to assume that first-person experiences in individ-uals of a given species have some things in common. Soit is no surprise that, while I can at least surmise as ahuman what it is like to be you as another human, itis not possible to be nearly as certain in trying to imaginewhat it is like to be a bat.

I shall indulge later in an exercise to see how our


model of primary consciousness might give rise to a senseof scene even in a bat. But first it is useful to point outthat we already have ample evidence from neuroscienceto suggest why different qualia have different feels. Theneural structures and dynamics underlying vision are dis-tinct from those of smell, and those for touch differ fromthose of hearing, and so on. Although no scientific de-scription of these pathways and their activities can giverise to a specific quale in the reader’s mind, if we assumethat he or she has an adequately equipped nervous sys-tem, he or she can relate such a description to a first-person experience. No matter what structure underlies aquale, it can be discriminated from others. One mightsay: “If it weren’t this way, it would be that way.” Thefact that it requires a particular body and a particularbrain in a particular environment is no great hindranceto a general analysis of the origin of different qualia.

According to the extended TNGS, qualia are high-order discriminations in a complex domain. The experi-ence of a conscious scene as unitary suggests the viewthat all conscious experiences are qualia. In this view,the separation of qualia into single, narrow feelings suchas red, warm, and so forth, while thinkable and verballydescribable, does not constitute a full recognition of thediscriminations involved. For example, we can, as scien-tists, describe color experiences in terms of a variety ofdefining properties, such as the spectral characteristicsof the three retinal pigments and the neural responsesof a given visual system. We can then plot the various


properties of particular experienced colors each as dis-tinct points in a three-dimensional space. But how dowe know that the color qualia are in fact what they are,except in the higher-dimensional space that also mapsthe various other qualia, thus allowing these mutual dis-tinctions? It is in fact the ability to discriminate refineddifferences of, say, hot and cold in the presence of myr-iad other qualities such as color in a unitary scene thatdistinguishes a conscious discrimination from the hot-cold distinction made, for example, by a thermostat. Tobe conscious is to be able to make such decisions basedon multidimensional discriminations or distinctions.

The richness of these differentiable states of con-sciousness and the unitary nature of each state do not,at first glance, seem compatible. To show that they arecompatible, it will suffice to provide a satisfactory ac-count of how the organization of the nervous systemcan give rise to these properties. These properties areprecisely those found in complex systems; therefore Iwill begin by providing a brief description of some prop-erties of complex systems. A complex system is a systemthat consists of a variety of smaller parts, each one ofwhich may be functionally segregated. As these hetero-geneous parts interact in various combinations, there isa tendency to give rise to system properties that are moreintegrated. My colleagues and I have described such sys-tems formally. Here, I will give a qualitative accountthat I believe will serve our purposes. The terms andmathematical measures used for characterizing a complex


system are borrowed from statistical information theory,but the premises of that theory are not accepted by ouranalysis. Such terms include “independence,” “en-tropy,” “mutual information,” and “integration.” I shallgive a few examples of their use, which I believe willclarify their meaning without going into mathematicaldetail. My goal is to show how a complex system candisplay integration of its parts but at the same time havemany differentiated states combining the properties ofthese parts.

Let me begin by presenting two extreme examplesof systems that are not complex (Figure 8). An ideal gaswith particles randomly colliding in elastic collisions isnot a complex system. Each particle is independent (itdoes not stick to the others), and there is no gain or lossof information (“mutual information”) exchanged insuch a collision. At the other extreme, a perfect crystalis not complex. In the case of such perfect regularity,there is a high degree of integration and mutual infor-mation among the units. However, once one knows theso-called space group and the contents of one of the unitcells of which the crystal is composed, no new informa-tion is gained in passing to any of the other unit cells.

Now let us consider a complex system. How is itboth integrated and differentiated at the same time? Wecan express the integration of a system in terms of itsso-called informational entropy. This is the amount ofinformation it would take to distinguish this systemfrom all possible similar systems made of the same com-


Figure 8. The contrast between the brain and twohypothetical systems—an ideal gas and a perfect crystal—which have much lower complexity. As a complex system,the brain has small, relatively independent parts that areheterogeneous in structure and function. As they connectby means of various kinds of neuroanatomy, they tend to

become integrated across a large number of statesgenerated by the functional connectivity within thatanatomy. In the case of the gas, no such integration

occurs whereas with the crystal there is high integrationbut no variety.

ponents, taking into account their relative probabilitiesof occurrence. Integration is the sum of the entropiesof each of the parts of a system minus the entropy ofthe system as a whole. In the case of an ideal gas, thisdifference is zero—putting separate bits of the gas to-gether to make a larger volume does not add any newinformation. But as parts of a system interact and sharemutual information (as in a crystal), the entropy of thesystem is less than the sum of the entropies of its parts,and integration takes on a positive value. In a perfectcrystal that value is as high as it can be.


We are now in a position to say in more precisefashion what characterizes a complex system and applythis characterization to the nervous system. Unlike afully integrated system, such as a perfect crystal, whenone considers smaller and smaller parts of a complexsystem, they deviate from a linear dependence of inte-gration and show more independence. But in the otherdirection, as larger and larger subsets consisting of suchinteracting parts are considered, they approach closer tothe limit set by a totally integrated system. This is justthe property seen in interacting networks of the brain.They exhibit functional segregation (cortical area V1 fororientation, V4 for color, V5 for object motion, andso on) but, through binding via reentry, they becomeintegrated—that is, they exhibit more unitary propertiesas they are linked together.

We may now apply these ideas to the thalamocorti-cal system in order to reveal a mechanistic neural basisfor the unitary yet differentiated properties of a con-scious scene or qualia space, the space representing allthe different qualia. But before we do, two matters mustbe considered in addition to what we have said so far.The first is that the thalamocortical system is dynamic.As a result of its enormous numbers of neuronal connec-tions, the reentrant interactions of its excitatory andinhibitory neurons as well as the gating effects of thereticular nucleus and subcortical value systems, thethalamocortical system shows rapid changes in its func-tional connectivity over fractions of a second. The sec-


ond matter concerns the relatively large number of inter-nal interactions in that system compared with thenumber of interactions it has with subcortical systems,such as the basal ganglia, that mediate nonconscious ac-tivities of the brain. It appears that the dynamic re-entrant thalamocortical system speaks mainly to itself.This defines what we call a functional cluster: most ofits neural transactions occur within the thalamus andcortex themselves, and only a relatively few transactionsoccur with other parts of the brain. As we shall see later,this is an important property that serves to distinguishthe neuronal activities subserving consciousness fromthose that do not.

This functional cluster with its myriad of dynamicreentrant interactions, occurring mainly, but not en-tirely, in the thalamocortical system, has been called thedynamic core (Figure 9). The dynamic core, with itsmillisecond-to-millisecond utilization of an extraordi-nary complex of neural circuits, is precisely the kind ofcomplex neural organization necessary for the unitaryyet differentiable properties of the conscious process. Ithas the reentrant structure capable of integrating orbinding the activities of the various thalamic nuclei andthe functionally segregated cortical regions to producea unified scene. Through such interactions, the dynamiccore relates value-category memory to perceptual cate-gorization. In addition, it serves to connect conceptualand memory maps to each other. Changes of state inthe dynamic core in response to signals from within and


without engage new sets of dynamic functionally segre-gated circuits in short times, and this property accountsfor the differentiation of successive scenes constitutingthe conscious state. Above all, because of the degeneracyand associative properties of its component circuits andneuronal groups, the activity of the core enables con-scious animals to carry out high-order discriminations.Qualia are these discriminations. The great variety of

Figure 9. The dynamic core. The thalamocortical system,which gives rise to the dynamic core, is represented by a

fine meshwork of cortical and thalamic areas and reentrantconnections. The core is composed of a functional clusterthat speaks mainly to itself through an enormous complex

of signals that fluctuate in time across the reentrantmeshwork. Responses triggered by the reentrant dynamic

core can also stimulate nonconscious responses. Thesetravel along parallel, polysynaptic, one-directional pathwaysthat leave the cortex, reach the various components of the

basal ganglia and certain thalamic nuclei, and finallyreturn to the cortex (as in Figure 3, middle diagram). Inthis way, responses subserving consciousness can connectto activity patterns in nonconscious areas, served mainlybut not exclusively by the basal ganglia. The ganglia andthe thalamus have been displaced and enlarged for clarity.Areas of the cortex not in the core at some particular timecan also interact with such nonconscious activity patterns.At some subsequent time hundreds of milliseconds later,

however, neuronal groups from these areas can participatein the core.


discriminations emerges because the dynamic core is acomplex system that can maintain functionally segre-gated parts while integrating the activity of these partsover time in a very rich combinatorial fashion. Eachtransient set of degenerate core circuits may underlie ascene, to be replaced in short time periods by anotherset, yielding a changed scene. Of course, this whole pic-ture is consonant with the TNGS and its notion of thebrain as a selectional system, one reflecting both con-stancy and variation.

In a developed individual, the range of integrationsover the entire qualia space can be enlarged by experi-ence or dynamically diminished by attention. Both pro-cesses are important in conscious planning. Think, forexample, of the subtle changes that occur in an oeno-phile after more and more refined discussions duringmultiple tastings of various categories of wine. But howcan such differential discriminations be made by justthat one person, that single self ? Can the workings ofthe reentrant dynamic core account for the fact that theconscious scene belongs to a subjective self ?

We will consider this at length when we discusshigher-order consciousness. Here I can provide a shortaccount of the changes that are likely to occur duringdevelopment and early experience. The earliest discrimi-nations of consciousness must concern perceptual cate-gorizations related to the body itself. These are mediatedthrough signals from structures in the brainstem and invarious value systems that map the state of the body (see


Figure 7). As I mentioned before, the signals from these“self” systems report the relation of the body to boththe inside and the outside environments. Such signalsinclude so-called proprioceptive, kinesthetic or somato-sensory, and autonomic components. These compo-nents, which signal, respectively, the position of thebody, the action of muscles and joints, and the regula-tion of the internal environment, affect almost every as-pect of our being. They regulate bodily functions ofwhich we, as mature individuals, are only dimly aware.Like the value systems that signal the salience of variousinternal and external events, these components are atthe deep center of conscious experience. The earlybodily-based consciousness of self (reinforced even byearly fetal movements) is likely to provide the initialguidelines of our qualia space, out of which all subse-quent memories, based on signals from the world (“non-self”), are elaborated. Thus, even before higher-orderconsciousness appears, a bodily-based neural referencespace or body-centered scene will be built up. An animalor a newborn baby will experience a scene in referenceto a self but will have no nameable self that is differenti-able from within. Such a nameable self emerges in hu-mans as higher-order consciousness develops during theelaboration of semantic and linguistic capabilities andsocial interactions. Qualia can then be named and ex-plicitly distinguished. But even before that, qualia arealready discernible and almost certainly referenced tothe ongoing categorization of the self by primary con-


sciousness. In the complex system underlying such con-sciousness, there are already hosts of qualia that consistof all the conscious states that can be discriminated. Thedynamic core, whose activities are enriched throughlearning, continues throughout life to be influenced bynew processes of categorization connected to whatmight be termed the bodily self. It is important to un-derstand, however, that the functional cluster compris-ing the core should not be identified at any one timewith all of the cortex or thalamus, parts of which arecontinually interacting with nonconscious regions of thebrain.

There remains the question of how the self is actu-ally aware of an ongoing scene. We must confront thedifference between the first-person experience and thethird-person description of the neural substate that un-derlies that experience. Perhaps it may be useful to imag-ine a demon or homunculoid observer whose task is toconfront what it is like to have that experience by inter-preting the metastable core states. Imagine that such anobserver is present in the brain and able to witness andmathematically interpret the myriad neural workings ofvalue-category memory in the dynamic core of a con-scious individual in a given animal species. Such a mem-ory system is already based on species-specific categoriesrelated to past perceptual experience. Now also imaginethat self-categories are at the forefront, even though theyare mixed with those perceptual categories that are re-lated to nonself. As ongoing activities of the reentrant


dynamic core lead to the creation of a new scene, ourhomunculoid can witness the neural activity responsiblefor creating that scene and also notice that the “self,”dynamically and continually constructed from bodilycues, can also be related to that scene. Even with thosecapabilities, however, this imaginary homunculoid doesnot and cannot perceive or control the high-level dis-criminations that underlie the conscious activities of theindividual in question. Nor can such a demon experi-ence the qualia that accompany these activities.

This somewhat bizarre construction suggests thateven with his analytical skills, the homunculoid couldnever come to know what it is like to be a conscioushuman. We inserted him into the brain and gave himthe capacity to read the core in an attempt to understandthe privileged nature of conscious experience. Observa-tions from without, or even from within, by a demonthat does not have the animal’s body cannot fully recap-ture the content of that privacy. But observing how cate-gorical memory referred to self can deal with new cate-gories to make a scene may be heuristically useful inimagining how the gap between subjective awarenessand neural action can be bridged. The homunculoid, ofcourse, does not exist. Indeed, the very imagining ofsuch a construct forces us to confront a central issue.This issue concerns the causal efficacy of consciousness,to which we now turn.


Consciousness andCausation

T H E P H E N O M E N A L T R A N S F O R M

e now come to the crux of our theory of con-sciousness. At this point in our account, we must con-front two questions, introduced earlier, both related tocausation. The first is: How is the conscious process en-tailed by neural processes? We have in a sense alreadyanswered this question, but the answer must be re-formulated to confront the second question, which is:Is consciousness itself causal?

76

Our previous account suggested that consciousprocesses arise from the enormous numbers of reentrantinteractions between value-category memory systemsthat are largely present in the more anterior parts of thethalamocortical system and the more posterior systemsthat carry out perceptual categorization. Through thecomplex shifting states of the dynamic core, these inter-actions underlie the unitary property of conscious states,as well as the shifting diversity of these states over time.Because the earliest interactions involve bodily inputsfrom centers of the brain concerned with value systems,motor areas, and regions involved in emotional re-sponses, the core processes are always centered arounda self that serves as a reference for memory. In primaryconsciousness, this self exists in a remembered present,reflecting the integration of a scene around a small inter-val of time present. While an animal having this primaryconsciousness has a long-term memory of past events,it has no extensive ability to deal explicitly with the con-cept of a past or a future. Nevertheless, it can carry outa vast number of conscious discriminations, discrimina-tions that are experienced as qualia. Only with the evo-lution of higher-order consciousness based on semanticcapabilities do explicit concepts of self, of past, and offuture emerge.

This account implies that the fundamental neuralactivity of the reentrant dynamic core converts the sig-nals from the world and the brain into a “phenomenaltransform”—into what it is like to be that conscious

c o n s c i o u s n e s s a n d c a u s a t i o n77

animal, to have its qualia. The existence of such a trans-form (our experience of qualia) reflects the ability tomake high-order distinctions or discriminations thatwould not be possible without the neural activity of thecore. Our thesis has been that the phenomenal trans-form, the set of discriminations, is entailed by that neu-ral activity. It is not caused by that activity but it is,rather, a simultaneous property of that activity.

This brings us directly to the second question. Isthe phenomenal transform itself causal? This question ispivotal, not only in considering how conscious acts occurbut also in addressing whether consciousness arose in evo-lution as an efficacious or adaptive process. To explorethis issue in a direct fashion, let us call the phenomenaltransform and its processes C. Call the underlying neuralcore processes C′. Both C and C′ could be indexed(C′0, C0; C′1, C1; C′2, C2; C′3, C3; and so forth) to indicatetheir successive states in time, but for now let us considerthem without addressing the temporal issue. We havepointed out that C is a process, not a thing, that it reflectshigher-order discriminations, and that it does not occurin the absence of C′. But, given the laws of physics, Citself cannot be causal; it reflects a relationship and cannotexert a physical force either directly or through field prop-erties. It is entailed by C′, however, and the detailed dis-criminatory activity of C′ is causal.

That is, although C accompanies C′, it is C′ thatis causal of other neural events and certain bodily ac-tions. The world is causally closed—no spooks or spirits


Figure 10. Causal chains in the world, body, and brainaffect the reentrant dynamic core. Core activities (C′) in

turn affect further neural events and actions. Coreprocesses confer the ability to make high-order

distinctions. The entailed phenomenal transform (C) withits qualia consists of those distinctions.

are present—and occurrences in the world can only re-spond to the neural events constituting C′ (Figure 10).

Consciousness C as a property of C′ is a reflectionof the capacity to make refined discriminations in a mul-tidimensional qualia space. This phenomenal transform,reflecting events in that space, is a reliable indicator ofthe underlying causal C′ events. The consequence of thisline of reasoning is that evolution selected C′ (underlainby the neural activities of the dynamic core) for the effi-cacy in planning conferred by its activity. At the sametime, however, such C′ activity entailed corresponding C


states. Indeed, there is no other way for an individualanimal to directly experience the effects of C′. The phe-nomenal transform provides an integrated scene that re-flects discriminations made possible by C′ activity andthus provides a coherent and reliable indicator to the in-dividual of the causal states underlying his consciousness.

This entailment of C by C′ also provides a cogentmeans for communication of C′ states to other individu-als. Even that communication has C′ as its causal vehi-cle. The relationship of entailment between C and C′implies that the so-called zombie argument of philoso-phers is logically impossible. That argument asserts thata zombie (an individual having C′ but without a phe-nomenal transform C) could carry out operations iden-tical to those of an individual with C. So, for example,without feelings, qualia, emotions, or a scene, a zombieart critic could, according to the argument, make identi-cal judgments about the superiority of one painting overanother to those made by a human art critic puttingforth the same judgments while experiencing C. Theargument we are making here implies, however, that ifC′ did not entail C, it could not have identical effects.The zombie would not know what it is like to be a con-scious human and could not carry out the necessary dis-criminations in a fashion identical to a human. More-over, being nonconscious, it could not be conscious ofbeing conscious. To have C′ as a result of core activitiesis to have C as a reliable property.

How might the C′–C relationship have evolved?


I have already considered the necessary development ofreentrant connections between brain regions carryingout perceptual categorization and value-category mem-ory. Here I want briefly to speculate on the origin ofthe relationship of entailment between C′ and C. It isreasonable to assume that the development of the abilityto carry out refined distinctions conferred by the dy-namic core would have selective advantage. The corecould conceivably have evolved even in species withoutextensive communicative abilities. I find it more attrac-tive, however, to consider that, in animal species inwhich rich communication of emotional states led toenhanced fitness, it would have been advantageous toconnect the ability (C′) to make refined distinctionswith the communication of these distinctions. Animalsso evolved would communicate efficacious C′ states interms of C. C, after all, is the only information availablethat reflects C′ states to each animal and to others. Aslong as C states reflect C′ states reliably, the fact thatthe world is causally closed and that only C′ is causalwould not undermine the role of C as a vehicle of com-munication.

The fact that the world is causally closed has beennoted by certain philosophers of mind, notably JaegwonKim. Following another philosopher, Donald David-son, Kim has proposed that a C state as a psychologicalstate is “supervenient,” or dependent on a physical state(in our terms, C′) that is causal. In early work, he hasdescribed all causal relations involving psychological


events as epiphenomenal supervenient causal relations.Presumably this refers to C′ as causal since “epiphenom-enal” means causally impotent. Although these notionsare roughly in accord with our account, I would notdesignate any mental event as directly causal, for it is arelationship and cannot exert a physical force. But theneural firings in C′ can do so, for example, by activatingmuscles. By providing a description of how C dependson C′ in a specific neural model we can go beyond anabstract statement about the dependence of C on C′.

In general, I have been in accord with, and eveninspired by, the views of William James on conscious-ness. But I do differ from him in interpreting the rela-tionship between consciousness and causality. In ThePrinciples of Psychology, James quotes T. H. Huxley,Darwin’s bulldog:

The consciousness of brutes would appear to berelated to the mechanism of their body simply asa collateral product of its working, and to be ascompletely without any power of modifying thatworking as the steam-whistle which accompaniesthe work of a locomotive engine is without influ-ence on machinery. Their volition, if they haveany, is an emotion indicative of physical changes,not a cause of such changes . . . to the best of myjudgement, the argumentation which applies tobrutes holds equally good of men; and, therefore,that all states of consciousness in us, as in them,


are immediately caused by molecular changes ofthe brain-substance. It seems to me that in men,as in brutes, there is no proof that any state ofconsciousness is the cause of change in the motionof the matter of the organism. If these positionsare well based, it follows that our mental condi-tions are simply the symbols in consciousness ofthe changes which take place automatically in theorganisms; and that, to take an extreme illustra-tion, the feeling we call volition is not the causeof a voluntary act, but the symbol of that state ofthe brain which is the immediate cause of thatact. We are conscious automata.

James takes issue with this view, which he calls the“Automaton-Theory.” He grasps its import and evenadds his own metaphors, saying: “So the melody floatsfrom the harp-string, but neither checks nor quickensits vibration; so the shadow runs alongside the pedes-trian, but in no way influences his steps.” Then hemounts a counterargument, insisting that the particularsof the distribution of consciousness point to its beingefficacious. His argument rests on three legs. First, heargues that consciousness is a selecting agency. Next heargues that the cerebral cortex is inherently unstable andthat this apparent defect can be corrected by conscious-ness, which stabilizes the cortex by being a “fighter forends,” reinforcing activity favorable for the organismand repressing unfavorable activity. Third, James argues


from the fact that pleasures are associated with beneficialexperiences and pains with detrimental ones. If pleasureand pains had no efficacy, he did not see why the reverse(pain, beneficial; pleasure, detrimental) could not betrue were the automaton theory to be correct. Con-sciousness, it seemed to James, is added in evolution “forthe sake of steering a nervous system grown too complexto regulate itself.” James was too honest a partisan notto point out an essential mystery entailed by his posi-tion: “How such reaction of the consciousness upon the[nerve] currents may occur must remain at present un-solved.” It is noteworthy that, more recently, anothergifted scientist, Roger Sperry, has taken the position thatconsciousness can actually affect neuronal firing.

Obviously, I have taken a contrary position: Noth-ing prevents us from espousing the view that all of thepoints made by James can be answered by the appro-priate evolution of C′ states along with their correspond-ing C states. Provided that a suitable mechanism of con-sciousness—that it arises from the activity of thereentrant dynamic core—is provided, there is no prob-lem concerning effects on “nerve currents.”

If I disagree with James, I must also take issue withHuxley: we are not automata. The TNGS, with its firmgrounding in population and selectionist thinking, re-jects the notion that we are machines or, more precisely,that we are Turing machines. Indeed, the variability ofconsciousness, which arises from the nature of the dy-namic core, is not a defect. This is so because the vari-


ability is accompanied by integrative activity and selec-tion. The very richness of core states provides thegrounds for new matches to the vicissitudes of the envi-ronment. Those matches are stabilized through theworkings of the brain as a complex system.

What is unusual about the position we are takingis not that C is an epiphenomenon or, if it is, that itposes a paradox. In fact, it does not. The unusual aspectof our view of causation is that C states, even while notdirectly causal, reliably reflect the incredibly refined dis-criminatory capacity of C′ states. C states or qualia arethe discriminations entailed by C′ states. This is the ba-sic feature of the conscious activity that results from thereentrant interactions of the dynamic core.

The tight relationship of entailment between C′and C necessarily involves a first-person experience. Anyalternative third-person assessment of C′ consequences(as with our homunculoid) would require an extraordi-narily rapid mathematical synthesis of an individual’simmediate core state, and a means for connecting thissynthesis of extraordinarily complex events to subse-quent events in the core. Clearly, evolution, powerfulas it is, could not assure such capabilities. Moreover, toallow an effective measure of causal consequences, suchcapabilities would also require a more or less completeknowledge of each individual’s prior value-category his-tory. Given the existence of novelty and the selectionalnature of neural events, a synthesis of this type couldnot be carried out by a computer, no matter how power-


ful. Even our bizarre homunculoid could not experiencethe processes we presumed it could follow.

When talking to each other as if our C states arecausal, we do not have to be concerned about the partic-ular C′ state that is the true cause of our exchange. Therelation of entailment that makes C a property of C′ isan accurate track of the relationship of C′ to causal effi-cacy. Although at first glance it may seem somewhateerie that all of our transactions, first- and third-person,rely on neural events, there is in fact no contradictionimplied. The only contradictions that might arise derivefrom the contrary assumptions: that C′ states can leadto identical effects without entailing C, that C can existwithout C′, or that C is itself causal.

The phenomenal transform is an elegant means ofconveying the integrated states of C′ on a first-personbasis. There is no other way to directly experience theseneural events. Even in the interchange between two con-scious humans, the phenomenal transform provides anindicator of causal relations without being causal itself.The subjective state reflects the ongoing properties ofthe neural states of the core. It is qualia space itself—consciousness in all its richness.


The Conscious and theNonconscious

A U T O M A T I C I T Y A N D A T T E N T I O N

e are all familiar with habit and with auto-matic activities, such as riding a bike, based on previousconscious acts of learning. We are also familiar with thevarious levels of our conscious attentive acts. Theserange from a kind of free-floating “rest state” of diffuseattention and awareness to highly focal attention to oneidea, image, or thought. All of these phenomena are re-lated in one way or another to the functioning of sub-cortical structures that collaborate with the dynamic

87

thalamocortical core. These subcortical structures—thebasal ganglia, the cerebellum, and the hippocampus—were described earlier. I have called these the organs ofsuccession because of their relation to movement andtime.

There is no doubt that the basal ganglia and cere-bellum are important in the initiation and control ofmovement. As I have already mentioned, the hippocam-pus is concerned with the conversion of short-termmemory into long-term memory by interacting with thecerebral cortex. After the bilateral removal of the hippo-campal formation, episodic memory can no longer beestablished, although all episodic memories previous tothe lesion remain intact.

In considering automaticity and attention, I shallfocus mainly on the transactions between the basal gan-glia and the cerebral cortex. These transactions connectnonconscious functions to conscious ones. To revealthose transactions, I must make another excursion intoneuroanatomy. This excursion, which requires consider-ation of a small thicket of Latin names, will reveal strongdifferences between the organization of the thalamocor-tical core and the system of basal ganglia (see Figure 3).But the names are not as important as the anatomicalconnections to which they give rise.

The basal ganglia are five deep nuclei that are atthe center of the brain (Figure 11). They receive connec-tions from the cerebral cortex and then send projectionsto the cortex by way of the thalamus. Their connections

t h e c o n s c i o u s a n d t h e n o n c o n s c i o u s88

to the cortex are topographically organized (that is, likea map) in both directions. They also connect to eachother in a series of polysynaptic loops.

A major portion of the basal ganglia, constitutinginput nuclei from the cortex, is the so-called striatum,which consists of the caudate nucleus and putamen. Theremaining nuclei are the globus pallidus, the substantianigra, and the subthalamic nucleus. The globus pallidusand one part of the substantia nigra make up the majoroutput nuclei projecting to the thalamus. Their outputmay be looked upon in turn as the input to the dynamicthalamocortical core. In addition to the input to the stri-atum by the cerebral cortex, the intralaminar nuclei ofthe thalamus also project to the striatum. It is particu-larly important to note that the basal ganglia receive in-puts from practically all regions of the cortex. This isin sharp contrast to the cerebellum, which gets inputfrom a more restricted sensorimotor portion of the cor-tex. While the cerebellum links to the motor and premo-tor regions of the cortex, the basal ganglia project as wellto the prefrontal cortex and to the so-called associationareas, the activity of which helps to weigh decisions re-lated to action.

According to classical accounts, there are two maincircuits through which the basal ganglia exert their ef-fects. We can best describe the effects of the basal gangliaby considering only motor activities for now; the samemechanisms apply to the other circuits engaged by theganglia. The so-called direct pathway receives excitatory


(glutamatergic) inputs from the cortex to the striatum.The striatum then projects to the internal segment ofthe globus pallidus and to the so-called pars reticulataof the substantia nigra. Both of these project in turn tothe thalamus and then back to the cortex. The indirectpathway follows a different route. It goes from the stria-tum to the external segment of the globus pallidus,which then sends fibers on to the subthalamic nucleus.This nucleus then projects back to the pallidus and sub-stantia nigra.

The outputs from the striatum are inhibitory. Ex-citation of striatal neurons by the cortex inhibits the in-

Figure 11. The motor circuit of the basal ganglia,illustrating the polysynaptic loops carrying inhibitory anddisinhibitory signals via the thalamus. This circuit of the

basal ganglia is a subcortical feedback loop from the motorand somatosensory areas of the cortex, traveling through

segregated portions of the basal ganglia and thalamus, andthence back to the premotor cortex, the supplementary

motor area, and the motor cortex. (The basal ganglia andthalamus have been displaced and enlarged for clarity.)

The basal ganglia are involved in regulating motorprograms and plans and also appear to be concerned withvarious aspects of attention. This diagram does not showthe more extensive and similarly organized projections ofthe basal ganglia to other brain areas, such as parts of theforebrain (frontal and parietal cortex), that are particularly

important for attention.


hibitory cells in the output nuclei of the basal ganglia.This releases (disinhibits) the thalamic cells and can leadto movement as a result of the stimulation of premotorand motor areas of the cerebral cortex. In contrast, theindirect pathway exerts its effects when corticostriatalinput inhibits the external segment of the globus pal-lidus. This results in disinhibition of the subthalamicnucleus, which then excites the output nuclei of thebasal ganglia via the transmitter glutamate. The resultis inhibition of the thalamus and diminished excitationof the motor areas.

A key modulation of both of these pathways comesfrom the action of the neurotransmitter dopamine,which mediates the projections from the substantia ni-gra. Dopamine excites the direct pathway but inhibitsthe indirect pathway. The net result in both cases is toenhance movement.

As this account indicates, the basal ganglia and thecerebral cortex are very differently organized. It is clearthat the motor circuits of the basal ganglia modulatemovement by enhancing some cortical responses andsuppressing others. Lesions of the dopaminergic projec-tion from the substantia nigra lead, for example, to Par-kinson’s disease. In this disorder, there are difficultiesin initiating movements, slowness in executing move-ments, tremors, and rigidity. But the involvement of thebasal ganglia in carrying out motor programs is probablynot the only effect these nuclei have on cortical func-tioning. There is evidence that patients with Parkinson’s


disease can have cognitive defects and overall slowing ofthought. There is also evidence that the basal gangliaare involved in defects and repetitive actions connectedwith obsessive-compulsive disorders. Moreover, in he-reditary Huntington’s disease, resulting from loss ofcholinergic cells and GABAergic cells in the striatum,there are severe cognitive defects. These include absent-mindedness and, finally, dementia accompanied by se-vere movement disorders. These severe cognitive defectsare probably related to the effects of the disease on theprojections of the basal ganglia to the prefrontal cortex.

A series of observations obtained by brain scanningtechniques is consistent with the hypothesis that theconnections between basal ganglia and cortex are in-volved in the execution of automatic motor programs.During conscious learning of tasks, a considerableamount of the cerebral cortex is engaged. With practice,conscious attention is not required, and acts become au-tomatic, as, for example, after learning to ride a bicycle.At such a point, brain scans show much less involvementof the cortex unless novelty is introduced, requiring fur-ther conscious attention. It is an attractive hypothesisthat collaborations between the cortex and basal gangliaset up the synaptic changes that lie behind such proce-dural learning. So, for example, practicing musical pas-sages will eventually result in the ability to “rattle themoff” without detailed attention. Later, two such learnedpassages may be joined by conscious efforts and furtherpractice, again to become automatic in execution. While


performing, a pianist playing a concerto with an orches-tra may execute passages without conscious attention,note by note, but simultaneously may plan consciouslyor think ahead about an upcoming musical phrase ortempo. The nonconscious portion of this execution is,by our hypothesis, governed largely by interactions be-tween the basal ganglia and the parts of the cortex notengaged in core activities.

The implication of this hypothesis is that certainportions of the cortex may be engaged in these interac-tions without being directly involved in the operation ofthe dynamic core. But when necessary, input and outputfrom the core can invoke learned responses involvingthose portions of cortex and basal ganglia, routines thatwere previously nonconscious.

In such interactions, attention can become in-volved to varying degrees, and it is likely that consciousattention is mediated by more than one mechanism. Forexample, in a “free-floating” or “rest” state of conscious-ness with little focal attention, it is reasonable to expectthat corticocortical reentry and shifting thalamocorticalreentry would provide sufficient bases. In more focal,but not highly exclusive, states of attention, gating bythe reticular nucleus of the specific nuclei of the thala-mus might come into play and restrict core activity. Andin highly focal attentive states, a reasonable surmise isthat the loops of interaction between the basal gangliaand the frontal and parietal cortices engaged in the coremay provide a central mechanism. This hypothesis as-


sumes that motor components of attention play an es-sential and even controlling role in imagined acts, butwithout engaging actual movements. The basal ganglioncircuitry is well suited for this. Unlike the circuitry ofthe cerebellum, it has no direct output connections tothe brainstem and spinal cord, but is connected to largeamounts of the cortex through the thalamus.

By this means, during highly focal attention,perceptual-motor loops and global mappings may serveto limit dynamic core states to the target of consciousattention. In this condition, it is as if the attending sub-ject is unconscious of all but the attended task. The in-hibitory loops of basal ganglion circuits and the abilityto modulate inhibition by balances between the directand indirect pathways would seem to be well adaptedto this mechanism.

The hypotheses put forth here are based on thenotion that the complex reentrant dynamics of thethalamocortical core can be influenced by nonconsciousbrain activity. I have not dealt with the Freudian uncon-scious and the notion of repression, which remains tosome extent a vexed subject. But it is conceivable thatthe modulation of value systems could provide a basisfor the selective inhibition of pathways related to partic-ular memories. For now, it is sufficient for my purposesto deal with the interaction between conscious states andthose which are necessarily nonconscious.

The anatomical excursions in this chapter werenecessary to reveal the differences between nonconscious


structures like the basal ganglia and cerebellum and thethalamocortical arrangements of the core. Core elementsare enormously reentrant. In contrast, basal ganglion el-ements are in long inhibitory loops. While the loci ofinteraction of such loops are richly distributed, in themain, the complexity of the dynamic core is far greaterthan that of the polysynaptic loops of the ganglia. As Ipointed out before, the core acts as a functional cluster,interacting largely with itself to yield conscious states.The ability to open up or restrict cortical–basal ganglioninteractions, thus modulating the content of conscious-ness, can nevertheless affect the range of the core clusterand alter attentional states.


Higher-OrderConsciousness and

Representation

y account so far has been focused mainlyon primary consciousness. As far as we are aware, ani-mals with only primary consciousness lack a sense of thepast, a concept of the future, and a socially defined andnameable self. Moreover, they are not conscious of beingconscious. Lacking these functions does not mean thatthey lack a self, that they lack imagery in the remem-bered present, or that they do not have long-term mem-ory. Within the attentive focus of consciousness in the

97

remembered present, they can even carry out plans andreact in terms of their past value-category memory.

So what is missing? According to the extendedTNGS, they have no semantic capabilities. They are notable to use symbols as tokens to lend meaning to actsand events and to reason about events not unfolding inthe present moment. This does not mean that to havehigher-order consciousness an animal needs to have alanguage. There is some evidence that such primates aschimpanzees have semantic abilities, but little or no syn-tactical ability, and thus that they lack true language.There is, nevertheless, evidence that they are able to rec-ognize mirror images of themselves and to reason aboutthe consequences of the actions of other chimpanzeesor of humans. Given this and their semantic capabilities,it is likely that they have a form of higher-order con-sciousness.

Our main reference species for higher-order con-sciousness is ourselves. We not only have biological indi-viduality but, in addition to having a self acting in theremembered present, we have higher-order conscious-ness and a socially and linguistically defined self. Weare conscious of being conscious, have explicit narrativeawareness of the past, and can construct scenarios in animagined future. We have a true language because wehave syntactic capabilities in addition to our phoneticand semantic capabilities. Given the acquisition and ac-cumulation of a lexicon, humans can use verbal tokensand symbols to divorce themselves from the remem-

h i g h e r - o r d e r c o n s c i o u s n e s s98

bered present by acts of attention. Of course, for higher-order consciousness to operate, primary consciousnessis still necessary, even if its immediate claims are tempo-rarily displaced by such acts of attention.

A number of interesting questions are promptedby these observations. One is related to the functioningof the hippocampus. This neural structure is necessaryfor episodic memory, the long-term memory of sequen-tial events, the brain’s “narrative.” As I have mentionedbefore, it is known that extirpation of the hippocampusbilaterally in adults eliminates the transformation ofshort-term consciousness and memory into long-termmemories of experience. After bilateral hippocampal re-moval, a patient retains episodic memories and narrativecapabilities for events up to the time of the removal. Butafter the operation, he or she cannot recall a sequence ofexperiences except for very short time periods. It is notknown whether being born without the hippocampusbilaterally would mean that an individual would lackconsciousness. But I would conjecture that, even if someform of primary consciousness were retained, it is likelythat higher-order consciousness would not develop.Higher-order consciousness rests in part on episodicmemory, and in the absence of such memory coherentsemantic activity would not be likely to develop.

The origin first of semantic abilities and then ofsyntactic abilities during evolution is a matter of debate.It is certainly simplistic to suggest that linguistic abilitiesemerged only with of the development of Broca’s and


Wernicke’s areas of the cerebral cortex. When theseareas in the cortex are damaged, it can lead to variousforms of the linguistic impairment known as aphasia.Novel subcortical structures, as well as the expansion ofthe prefrontal cortex, were also likely to have been in-volved in the evolution of regions making grammar andtherefore true language possible. Whatever the case, itseems likely that new reentrant pathways and circuitsappeared among these brain regions as a key basis forthe evolutionary emergence of semantic and, finally, lin-guistic ability during evolution. The appearance of suchpathways is thus critical for the development of higher-order consciousness (Figure 12).

As these capabilities develop in an individual, therange of conscious thought expands greatly. The brainis capable of going beyond the information given, as thepsychologist Jerome Bruner puts it. Reentrant interac-tions between maps mediating concepts, those mediat-ing linguistic tokens, and the nonconscious portions ofthe brain make it possible for consciousness to exploitmemory, even in the absence of new perceptual infor-mation. It must not be imagined, however, that linguis-tic performance will immediately guarantee the fullcapabilities of higher-order consciousness. During child-hood, increasing competence must develop and be coor-dinated with conceptual and memory systems before thefull flowering of higher-order consciousness.

Anthropologists have speculated about the devel-opment of language in hominines. Surely one cannot


Figure 12. Evolution of higher-order consciousness. A newreentrant loop appears in primates with semantic

capabilities and expands greatly during the evolution ofhominines with the emergence of language. The

acquisition of a new kind of memory, exploiting semanticcapabilities and ultimately true language with syntax leads

to a conceptual explosion. As a result, higher-orderconsciousness appears and concepts of the self, the past,

and the future become connected to primaryconsciousness. Consciousness of consciousness becomespossible. (See Figure 7 to relate this scheme to that for

primary consciousness.)

ignore the development of the vocal tract and spaceabove the larynx for speech, the motor regulation of res-piration during speech, and the bases for the auditorydiscrimination of coarticulated sounds. Much has beenmade also of the enormous and evolutionarily rapid de-velopment of the cerebral cortex in human evolution.The adoption of bipedal posture was almost certainly anecessary prerequisite for cortical enlargement, as it liftedconstraints on craniofacial morphology, thus providingaccommodation for a large cortex.

One wonders whether, in addition, bipedalism alsoallowed complete freedom of the forelimbs so that ges-tures assumed increased communicative significance. Insome deaf populations and certain deaf-mute individu-als, gestural communication without the syntax of truesign language has been observed. Such communicationby mime suggests that linguistic precursors may haveincluded gesturing as well as vocal signs. In freeing theupper extremities from brachiation (climbing or hang-ing) or walking, a whole precursor set involving the in-terpretation of gestures by the self and by others mayhave been opened up for early hominines. Whether in-fants who have learned to walk, and have their upperlimbs free, develop similar capabilities before the exer-cise of extensive speech acts is a question that remains.The acquisition of language may be enormously facili-tated by the development of conscious imagery relatedto movements and motor control. Almost certainly,concepts of objects, events, and succession must exist in


a child’s mind before the exercise of language. Accordingto these ideas, the sequences of actions of the free upperextremities may prepare the basal ganglion–corticalloops for the emergence of syntactical sequences, estab-lishing what might be called a protosyntax.

Clearly, one of the largest steps toward the acquisi-tion of true language is the realization that an arbitrarytoken—a gesture or a word—stands for a thing or anevent. When a sufficiently large lexicon of such tokensis subsequently accumulated, higher-order conscious-ness can greatly expand in range. Associations can bemade by metaphor, and with ongoing activity, earlymetaphor can be transformed into more precise catego-ries of intrapersonal and interpersonal experience. Thegift of narrative and an expanded sense of temporal suc-cession then follow. While the remembered present is,in fact, a reflection of true physical time, higher-orderconsciousness makes it possible to relate a socially con-structed self to past recollections and future imagina-tions. The Heraclitean illusion of a point in the presentmoving from the past into the future is constructed bythese means. This illusion, mixed with the sense of anarrative and metaphorical ability, elevates higher-orderconsciousness to new heights. Consideration of these ca-pacities turns our attention to the issue of representation“in the mind.”

Modern cognitive science rests heavily on notionsof mental representation and, in some sectors, on thenotion that the brain carries out “computations.” To


a certain extent, neuroscientists tend to use the terms“representation” or “encoding” somewhat differently torefer to the correlation or covariance of patterns of neu-ral firings with perceptual inputs or memory states. Un-less a distinction is made about the role of consciousnessin applying these terms, confusion can result.

As a term, “representation” has had a very liberalusage—it has been applied to images, gestures, lan-guage, and so on. In most cases, reference and meaningare attached to its use. But it is a common error toequate meaning and mental representation, for reasonsthat I shall consider in due course. In any event, it isdifficult to avoid the conclusion that consciousness andrepresentation are intimately related. While it is possiblefor a neurophysiologist to say that a pattern of firingthat is correlated with an input signal is a representation,this usage reflects a third-person point of view. As such,it does not encompass mental imagery, concepts, andthoughts, and certainly not the freight of intentional-ity—that is, beliefs, desires, and intentions.

The position I shall take here is that, although theconscious process involves representation, the neuralsubstrate of consciousness is nonrepresentational. Thishas the corollary that forms of representation occur inC, but do not compel the underlying C ′ states (see fig-ure 10). In this view, memory is nonrepresentationaland concepts are the outcome of the brain mapping itsown perceptual maps leading to generalities or “univer-sals.” While memory and concepts are, together with


value systems, necessary for meaning or semantic con-tent, they are not identical to that content.

The advantage of this position is that it does nottie the questions of meaning and reference to a one-to-one correspondence with either brain states or environ-mental states. At the same time, the enormous varietiesof “representations” can be accounted for in terms ofstates of primary and higher-order consciousness. Forexample, mental images arise in a primary-consciousscene largely by the same neural processes by which di-rect perceptual images arise. One relies on memory, theother on signals from without. Concepts, on the otherhand, need not rest on imagery but rather depend onglobal mappings and certain activities of motor systems,which need not necessarily engage the motor cortex, andtherefore do not lead to movement. At a higher level,cognition and intentionality are simply parts of the con-scious process that may or may not entail imagery.

This view rejects the notion of computation and theidea that there is a “language of thought.” Meaning isnot identical to “mental representation.” Instead, it arisesas a result of the play between value systems, varying envi-ronmental cues, learning, and nonrepresentational mem-ory. By its very nature, the conscious process embeds rep-resentation in a degenerate, context-dependent web: thereare many ways in which individual neural circuits, synap-tic populations, varying environmental signals, and previ-ous history can lead to the same meaning.

The problem of representation and intentionality


is related to the problem of explaining how higher-orderconsciousness itself arises. The essential issue to grasp isthat the reentrant circuitry underlying consciousness isenormously degenerate. There is no single circuit activ-ity or code that corresponds to a given conscious “repre-sentation.” A neuron may contribute to that “represen-tation” at one moment, and in the next have nocontribution to make. The same is true of context-dependent interactions with the environment. A shift ofcontext can change the qualia that are parts of a repre-sentation, or even recompose some qualia and still keepthat representation. In any event, there are other aspectsof qualia related to sensation that are not included inany representation.

The relationships to the processes of integration anddifferentiation in the complex dynamic core underlyingconsciousness follow directly from these notions. Corestates themselves do not “represent” a given image, con-cept, or scene in a one-to-one fashion. Instead, dependingon input, environment, body state, and other contexts,different core states can underlie a particular representa-tion. The interactions are relational and have the proper-ties of polymorphous sets. These are sets, like LudwigWittgenstein’s “games,” that are defined neither by singlynecessary nor by jointly sufficient conditions. If, for exam-ple, there are n different criteria for games today, any sub-set m of criteria, where m is much smaller than n, maysuffice to define a game or, in the case at hand, a subsetof core states underlying a particular representation.


While this view does not have the crisp propertiesof logical atomism or of computer models of the mind,it is in accord with a number of observations on lan-guage and reference. By pointing out that, for any repre-sentation, there can be many underlying neural statesand context-dependent signals, it takes account of thehistorical nature of conscious experience. Above all, it isin accord with the enormous complexity of the relationsunderlying any given “representation.” The problem ofexplaining the enormous variety of representations canbe resolved by considering how relations arise amongthe variety of conscious states and, above all, amongtheir enormously complex neural correlates in the core.

So far, I have said little about experiments designedto reveal the neural correlates of consciousness. Thecomplexities of representation provide an opportunityto discuss briefly one experiment of this kind. It wasdesigned to ask what happens when a person becomesaware of a perceptual object. The results show that eachindividual’s brain shows widespread reentrant interac-tions when that subject becomes aware of an object. Thedata also show that different individuals reporting simi-lar conscious responses have different individual pat-terns—that is, each case is distinct from the others.

The experiment employs a noninvasive techniquecalled magnetoencephalography, which measures mi-nute electrical currents that occur within as few as tenthousand neurons by detecting the magnetic fields asso-ciated with these currents. It does this by using special


devices that rely on superconductivity, the conductanceof electricity essentially without resistance at very lowtemperatures. In one version of the apparatus, 148 ofthese superconducting devices are distributed in a hel-met over the living subject’s head. The measurement offields is carried out in a shielded room to minimize out-side interference.

The actual experiment relies on a phenomenonknown as binocular rivalry. The subject, wearing glasseswith one red lens and one blue lens, stares at a screendisplaying vertical red bars crossed at right angles byhorizontal blue bars. As constructive as the eye and thebrain are, these disparate images cannot be reconciledor fused into one. Instead, the subject first sees verticalred bars and then, perhaps a few seconds later, sees hori-zontal blue bars. So it goes alternately; each time red isseen the subject presses a right-hand switch, and eachtime blue is seen the subject presses a left-hand switch.The switch responses are recorded simultaneously withthe magnetic responses of the brain.

The data are treated by mathematical techniquesthat eventually yield plots of the magnetic intensity overspecific areas of the cerebral cortex when the subject re-ports consciousness of the object and also when no suchreport is given. (Remember that the brain “sees” the redverticals and blue horizontals whether the subject re-ports being conscious of them or not.) Another mathe-matical technique can be used to measure whetherneuronal groups in distant parts of the brain fire syn-


chronously or not, thus testing for the appearance ofreentry.

The results are remarkable. When the subject is notconscious of any bars, the brain nonetheless responds ina swath going from the visual areas in the posterior partsof the cortex to the frontal areas related to so-calledhigher functions. But when the subject reports beingconscious of red or blue bars, the responses show veryspecific patterns. Some brain areas show decreases in theintensity of their activity, while, in others, activity in-creases; in general, however, awareness of a target pat-tern is accompanied by increases of 40 to 80 percent inbrain responses (Figure 13).

In one set of experiments, no two subjects hadidentical response patterns. An analysis of the synchro-nous firing of distant neurons in small intervals of timeshowed extensive evidence of reentrant interactions. Al-though each subject had a similar response to report (a“representation” of either blue horizontal or red verticalbars), the patterns recorded for each subject were indi-vidual and different from those of any other subject.Although this experiment has yet to be repeated for eachindividual over long periods of time, it is clear that,among different individuals, each “representation,”however similarly reported, was correlated with widelyvariant reentrant patterns.

The position and experiments I have just discussedhave several consequences. The first is that, while of im-mense importance, neurophysiological recording cannot


Figure 13. Coherence of neural processes underlyingconsciousness, measured using magnetoencephalographyduring binocular rivalry. The thin straight lines in whitegoing from front to back and side to side indicate the

increased synchronization that appears between distant anddistributed cortical regions when the subject is conscious

of a stimulus. The results are obtained by subtractingmeasurements when the subject was unconscious of the

stimulus from the measurements made when the subject isconscious of that stimulus. The top of the plot representsthe anterior regions of the cerebral cortex. The underlying

dark and light regions and the contour lines reflect theintensity of the brain responses measured by the magneticfields. These findings support the prediction made in the

extended TNGS that reentry is a central mechanismunderlying the conscious state.

[To view this image, refer to

the print version of this title.]

alone capture the richness of conscious representation.This must not be misunderstood—neurophysiologicalanalyses of covariant and causal relations in C′ are fun-damental. But given the complexity and degeneracy ofthe environmental and bodily input to the dynamiccore, there will be found no singular mapping for eachrepresentational state, just as there is none for similarqualia. While there are classes of neural states for a scene,it is hardly valuable to refer to such highly variable andcontext-dependent dynamic mappings in C′ by the term“representations.”

Another consequence of this view is that much ofcognitive psychology is ill-founded. There are no func-tional states that can be uniquely equated with definedor coded computational states in individual brains andno processes that can be equated with the execution ofalgorithms. Instead, there is an enormously rich set ofselectional repertoires of neuronal groups whose degen-erate responses can, by selection, accommodate theopen-ended richness of environmental input, individualhistory, and individual variation. Intentionality andwill, in this view, both depend on local contexts in theenvironment, the body, and the brain, but they can se-lectively arise only through such interactions, and notas precisely defined computations. Whether the self-directed movements of a fetus give rise to a “representa-tion” of the difference between its own efforts and im-posed external motions, or at the other extreme, whetherparticular Shakespearean metaphors and neologisms


strike adults as immediately meaningful, the view thata nonrepresentational process can give rise in many waysto conscious representations is in accord with a widerange of observations and possibilities, as well as withthe view that higher-order consciousness itself reacheda pinnacle with the acquisition of language.


Theory and the Propertiesof Consciousness

s it possible to summarize a theory of consciousnessin a short compass? I think it unlikely except if the sum-mary is addressed to those who have already taken the longexcursion. With that audience in mind, I shall try.

My first assumption has been that a biological the-ory of consciousness must rest on a global brain theory.This is the case because one must confront the enor-mous variability and individuality of higher brains andtheir dependence on value systems. The variability mustbe accounted for in terms of the principles of develop-

113

ment and evolution. My second assumption is based onthe recognition that principles of physics must be strictlyobeyed and that the world defined by physics is causallyclosed. No spooky forces that contravene thermody-namics can be included. My argument, which does notcontradict physics, has been that computer or machinemodels of the brain and mind do not work. Once weabandon logic and a clock, however, both of which arenecessary for the operation of digital computers, wemust provide an organizing principle for spatiotemporalordering and continuity in the brain. That principle isincorporated in the process of reentry.

All of these notions are subsumed within a selec-tional theory of brain function, the TNGS. In this the-ory, the variance and individuality of brains are notnoise. Instead, they are necessary contributors to neu-ronal repertoires made up of variant neuronal groups.Spatiotemporal coordination and synchrony are pro-vided by reentrant interactions among these repertoires,the composition of which is determined by develop-mental and experiential selection. Because of the degen-eracy of the neural circuits that arise dynamically as aresult of these selective processes, associative interactionsare guaranteed.

A theory of consciousness requires organizingprinciples for perceptual categorization and for value-category memory. According to the TNGS, perceptualcategorization takes place by means of global mappingsthat connect various modal maps through reentry and

t h e o r y a n d p r o p e r t i e s o f c o n s c i o u s n e s s114

also link them by non-reentrant connections to systemsof motor control. According to the theory, memory isnonrepresentational, and it is necessarily associative asa result of the interactions of degenerate networks.

With these premises of the TNGS in hand, an ex-tended theory can be formulated to deal with the neuralorigins of consciousness. Primary consciousness arises asa result of reentrant interactions between brain areas me-diating value-category memory and those mediatingperceptual categorization. A consequence of such inter-actions is the construction of a scene. The main sourcefor these transactions is the dynamic core, which is basedlargely in the thalamocortical system. The complexityof this core is enormous, but, as a result of dynamicreentry, certain of its metastable degenerate states canyield coherent outputs and the ability to distinguish var-ious modal combinations in a high-dimensional qualiaspace. That discriminatory capability within a unitaryscene is exactly what the process underlying primaryconsciousness is proposed to be. Qualia are the discrimi-nations entailed by that process. The individual, subjec-tive, and privileged properties of consciousness arise inpart because bodily systems are not only the earliest butalso the continually prevalent sources of perceptual cate-gorization and memory systems throughout life.

The extended TNGS claims to answer two ques-tions: (1) How do qualia arise in an individual? (2)What is causal about the neural and mental states oc-curring in an individual? The extended theory claims


that underlying any conscious state C is a set of neuralprocesses C′. Given the causally closed nature of theworld, it is C′ that is causal, and not C. But given thatit is a property entailed by C′, C is the only informationon C′ available to a subject (Figure 14).

It is essential to recognize that, in a strict sense,C′ does not cause C—there is no temporal lag in theexpression of C upon occurrences of C′. The proposedmechanism by which C′ gives rise to C as a propertydoes, however, include serial temporal changes that fol-low from neural dynamics. This mechanism also incor-porates the properties of those dynamics that followfrom the binding events among cortical maps that occurthrough the operation of reentry. These include filling-in (as experienced, for example, in our failure to see theblind spot), as well as a variety of gestalt phenomena.All of these characteristics are reflected in the unitarynature of each conscious scene. Nevertheless, each suchunitary scene is followed in short order by yet anotherscene and, indeed, by a host of differentiated core statesresulting from the bootstrapping in time between mem-ory and perception.

Higher-order consciousness, which allows its pos-sessor to be conscious of being conscious, to have asocially defined nameable self, and to have a conceptof the past and the future, arises by the evolution ofan additional reentrant capability. This occurs whenconcept-forming areas involved in primary conscious-ness are linked by reentrant circuits to areas mediating


Figure 14. Summary diagram of the causal interactions ofthe body, brain, and environment that give rise to primary

and higher-order consciousness. Causal events involvesignals from the self and world leading to action on the

world and interaction with other causal events (C′) in thedynamic core. The corresponding entailed properties (C)are the qualia, the high-order distinctions that constitutethe phenomenal transform, represented as the dotted areaon the left. Heavy arrows represent reentry; lighter arrows

represent causal loops. The abbreviations used are:1° C � primary consciousness; HOC � higher-order

consciousness; PC � perceptual categorization;VCM � value-category memory.

semantic capability. Present in higher primates, it reachesits most advanced expression in humans, who possesstrue linguistic capability. The ability to link the tokensin a lexicon by syntactical means greatly enhances therange of reentrant expression. While the higher-orderconsciousness that emerges still depends on primary con-sciousness, having tokens and means of this kind allowsan individual to become temporarily free from bondageto the remembered present.

This condensed summary is consistent with manyimportant properties related to the conscious state.Rather than expand this account to include them, it maybe more valuable to comment briefly on the testabilityof the extended TNGS and then to consider its explana-tory power. A biological theory of consciousness mustbe testable across a variety of levels ranging from themolecular to the behavioral. The most efficacious testswould focus first and foremost on the demonstration ofneural correlates of consciousness. As I discussed above,recent experiments at the Neurosciences Institute usingmagnetoencephalography to measure brain responses ofhuman subjects when they become aware of a visual ob-ject have revealed such correlates. Perhaps the most im-pressive feature of the experimental results was the find-ing of an increase in reentrant activity across wide areasof the cortex when the subjects reported their becomingaware of an object. Experiments in other laboratories arealso continually expanding our knowledge of the neuralcorrelates of consciousness.


In addition to testability, an adequate theory must,above all, lead to understanding and provide an explana-tion of known properties of the conscious state. Theseproperties fall into three feature categories, each ofwhich I shall consider in turn. First are those propertiesthat are shared by every conscious state, which I shallcall general or fundamental properties. Second, there areproperties related to the informational functions of con-sciousness. And, third, there are the subjective proper-ties—those related to feelings and to notions of the self.The various properties under each category are listed inTable 1.

My aim here is to show that the extended TNGS,as summarized above, is consistent with these propertiesand that it provides a biological basis for each of them.I shall not explicitly deal in this account with all ofthe details of states such as beliefs, desires, emotions,thoughts, and so on, which are derived by interactionsamong these properties. Once I have shown how thevarious properties can be accounted for, it will not bedifficult to show connections with those compositestates, which philosophers call propositional attitudes.

Consider first the general properties. Each con-scious state is unitary—it cannot be divided into sepa-rate parts as it is experienced. Instead, at any time, theconscious scene has unity. It is not possible, willfully orwith even a high degree of attention, to limit awarenessto one particular component of a scene to the exclusionof all others. Yet myriad conscious states and scenes can


Table 1. Features of conscious states

General1. Conscious states are unitary, integrated, and

constructed by the brain.2. They can be enormously diverse and

differentiated.3. They are temporally ordered, serial, and

changeable.4. They reflect binding of diverse modalities.5. They have constructive properties including

gestalt, closure, and phenomena of filling-in.

Informational1. They show intentionality with wide-ranging

contents.2. They have widespread access and associativity.3. They have center-periphery, surround, and fringe

aspects.4. They are subject to attentional modulation, from

focal to diffuse.

Subjective1. They reflect subjective feelings, qualia,

phenomenality, mood, pleasure, and unpleasure.2. They are concerned with situatedness and

placement in the world.3. They give rise to feelings of familiarity or its lack.

be experienced, and conscious states follow each otherin a temporal and serial order. The TNGS postulatesthat the reentrant dynamic core can give rise to preciselythese properties as a complex system: it has parts thatare functionally segregated, but within short time pe-riods they can become increasingly integrated. Corestates change from one to another within periods ofhundreds of milliseconds as different circuits are acti-vated by the environment, the body, and the brain itself.Only certain of these states are stable, and thus actuallybecome integrated, and it is this integration that givesrise to the unitary property of C. Because the core carriesout reentrant interactions between perceptual categori-cal input and value-category memory, both of which arecontinually changing, it also changes. The quasi-stablestates of the core represent the binding of various modal-ities in different cortical regions that occurs as a resultof reentrant interactions. The bound states arise fromdegenerate sets of circuits: contributions from all theneuronal groups within a given circuit are synchronous,but a similar output can emerge from different subsetsof circuits that follow each other in a serial and asyn-chronous fashion. The temporal properties of conscious-ness arise from these processes.

These neural activities account for the unitary, inte-grated, yet differentiated properties of C. But it is alsoimportant to point out that, according to the TNGS,the brain necessarily must be constructive. One aspect


of the integrative property of reentrant selectional net-works is the appearance of filling-in and gestalt prop-erties. Reentrant dynamics involve shifting dominancesbetween and among cortical maps. Because of this,and because the selectional units are groups of neuronswith differing properties, higher-order integrations canemerge in which one property can dominate or incorpo-rate another. This can be seen in various visual, auditory,or somatosensory illusions. Indeed, the deliberate designof illusory inputs by neuroscientists and psychologists toemphasize certain features, as compared to the more bal-anced habitual input stream of signals from the environ-ment, is likely to favor certain maps over others in areentrant economy. Consciousness is itself an internallyconstructed phenomenon. By this I mean that, althoughperceptual input is important initially, in relatively shortorder the brain can go beyond the information given, oreven (as in REM sleep) create conscious scenes withoutinput from or output to the external world. Those scenesare mediated by reentrant connections to those parts ofthe brain that can be engaged in perception and to thoseinvolved in concept formation.

These observations point to an important problem:how can C′ states that follow each other in a more or lesscontinuous fashion give rise to the more or less smoothsuccession of C states without stalling or interruption?I can only put forth a conjecture: linkage of C′ statesinvolves cyclic and concatenated reentrant interactions.Such “looped” and overlapping interactions would be


Figure 15. Hypothesis of reentrant dominance. Cyclic orconcatenated reentrant paths are more likely to sustain

ongoing activity than are linear paths. Dotted linesindicate a reduction or loss of reentrant signaling in linear

paths not connected to cycles.

favored over linearly connected dynamic circuits, evendegenerate ones (Figure 15). Although we presently lackthe means to test this idea, it is worth keeping in mind.

Certain objections have been raised to the hypoth-esis that a unitary “continuous” or grain-free scene inthe phenomenal transform might arise as a result of thediscontinuous firing of discrete neurons. But a littlethought about the overlapping distributions of the fir-ings of large numbers of neuronal groups in time andspace should dispel these concerns. Moreover, as a resultof the dynamics of reentry, the competitive dominanceof certain neuronal groups, and the contributions of cat-egorical memory, the brain generally tends to be con-


structive. Filling in of the blind spot, the phenomenaof apparent motion, and gestalt phenomena can all beexplained in terms of temporal synchrony in reentrantcircuits. The same is true of the sense of time, of succes-sion, and of duration. The reentrant brain combinesconcepts and percepts with memory and new input tomake a coherent picture at all costs.

Notwithstanding the unitary and constructive na-ture of the conscious state, the conscious scene still hasenormous richness of detail. Most of this is attributableto the actual richness of signals from the physical sur-round as they are filtered through each sensory modalityand modulated by memory. The actual content of a con-scious scene obviously depends on the presence or ab-sence of such a filter. A congenitally blind person lackingvisual cortical area V4 will never know what the colorred is. Yet because of the enormously parallel coincidentinput from the surround through hearing, touch, andmotion, a blind person can construct a “space” that canserve usefully to signify a number of functions and be-haviors. In general, the content of consciousness de-pends on whether certain cortical areas serving specificmodalities are functioning normally. A person’s phe-nomenal experience depends on these modalities and,as I have emphasized, the phenomenal aspect of thesemodalities cannot be reproduced by explanation. Evenan accurate theory of consciousness cannot provide theblind person with an experience of redness.

All these factors account for the “irreducibility” of


consciousness and the subjective state. While some feelit necessary to “reduce” conscious experience by identi-fying it with neural action, this reduction leads to a cate-gory error. The origin of qualia as properties of neuralprocesses having high-order discriminatory powers doesnot eliminate the subjective experience they represent.

With this account of the bases of the general prop-erties of consciousness in hand, we can turn to what Ihave called informational properties—those that pro-vide information reflecting input and output in C′. Thefirst of these is intentionality, a term proposed by thepsychologist Franz Brentano in the nineteenth century.It is the property by which consciousness is directed at,or is about, objects and states of affairs that are initiallyin the world. Not all forms of consciousness are inten-tional, and certain intentional states are not necessarilyconscious. In any case, the term is not coextensive with“intending”—intending is intentional, but “intention-ality” refers to a much wider range of referential states.The extended TNGS points out that the initial develop-ment of conscious states depends on interaction withperceptual categorization guided by value systems. Inso-far as this fundamental aspect of higher brain functiondepends on input from the world and the brain throughvarious modalities, it is not surprising that in both con-scious perceptual and memorial states, intentionality is acentral property. Clearly though, not all conscious states(mood, for example) are intentional.

Another aspect of the informational nature of states


underlying consciousness is their extraordinary associa-tivity and wide access to sensation, perception, memory,imagery, and various combinations of all these. The ex-tensive mapping of the reentrant dynamic core that iswidespread over the cortex is consistent with this prop-erty. In imagery, for example, reentry essentially engagesmore or less the same sets of pathways that would beoccupied in primary visual perception, along with otherassociative pathways. The property of associativityemerges from reentry and the degenerate interactions ofthe thalamocortical circuits that make up the core. Non-representational memory also has degenerate propertiesthat assure rich associations with a variety of circuits inaddition to those involved in any particular recall.

Surround effects and the fringe at the edge of theconscious scene necessarily accompany the operations ofthe complex functional cluster of core activity; these areinfluenced by nonconscious activities of the circuits ofthe basal ganglia. Given the rapid changeability andmetastability of core operations and the constructive na-ture of associative reentrant binding, fluctuations at theedge of the conscious scene would be expected. Take theeye and its movements, for example. The retina has acentral region of high discriminatory power (the fovea),and the eye itself moves in fast jumps called saccades. Invision, although a scene appears fairly uniform up to the“fringe,” central foveal discrimination is certainly moreprecise even though an individual is not aware of it. Sac-cades and smooth eye movements “paint” a more uni-


form, constructed scene as a result of the various trade-offs of brain states between precision and inclusivenessthat occur after the brain receives signals from the opticnerve. This is another example of constructive filling-in,which must have variation at its edges.

This brings us to the complex issue of attention,which I believe has multiple mechanisms. These rangefrom the relatively diffuse C states that result from theC′ states mediated by corticocortical interactions, tothose gated via the reticular nucleus of the thalamus, tothe most highly focal states of the core influenced bythe motor cortical circuitry of the basal ganglia. We arenot aware of those “blocked” motor states, but the the-ory suggests that it is the engagement of the core withcircuits lacking output to muscles that forms the basisof the most focal of attentive conscious states. In thesefocal states, the core is modulated to such a degree thatthe experience is as if one is anesthetized deeply to allaspects of an image, scene, or thought but the one thatis focally attended. The exact mechanism by which suchmodulation occurs is not known. One possibility is thatthe inhibitory output of global mappings to the thala-mus via the basal ganglia allows certain core responsesto occur at the expense of others. The details remain tobe worked out. In any event, it is likely that attentionis effected through a variety of different routes andmechanisms. I have already discussed the interactive as-pects of attentive learning and automaticity, which areconnected to the question of how automatic routines


previously learned by conscious attentive means are re-called and linked together consciously. The notion thatthis is achieved by interactions between the thalamocor-tical core and the basal ganglia (which may also engagethe cerebellum) is one that still requires testing.

I turn now to properties related to subjectivity. Thephenomenal transform of C′ to C through the earliestcategorical experiences of bodily perceptions is one ofthe major origins of the subjective sense and the notionof the self. I have made the statement that all consciousexperience partakes of multiple qualia and that a singlequale, say “red,” cannot be the sole aspect of consciousexperience. According to the extended TNGS, we expe-rience a multidimensional qualia space and conscious-ness reflects our ability to make high-order discrimina-tions, which are the qualia within that space. Obviously,different sensory modalities lead to different discrimina-tory capabilities. Their content depends particularlyupon the range of cortical interactions of the dynamiccore, which is modulated by attention. This is consistentboth with the unitary property of the conscious sceneand its differentiability.

The question arises as to the continued central roleof the self through contributions of the body, the envi-ronment, and memory. There are two contributionsthat appear to be fundamental. One is the phenomenaltransform acted upon by various modalities and, earlyon, by value systems, autonomic responses, and proprio-ception (see Figures 7 and 10). These systems, con-


cerned as they are with bodily regulation, must continueto operate throughout life in parallel with other inputsfrom sensory modalities.

The phenomenal contribution to self-reference isenhanced by a second contribution—the Piagetian no-tion of self—the distinction made between internallygenerated directed movement and motions inducedfrom outside sources. This discrimination may actuallyoriginate in utero during the late fetal stages, but cer-tainly occurs during early postnatal development. It pro-vides a reference for distinguishing self from nonselfthrough kinesthetic inputs that may act in addition to,and separately from, explicit sensory contributions toqualia space.

A third form of self-discrimination is likely toemerge later in development as a property of higher-orderconsciousness. This is the conscious process of individua-tion—the recognition of other selves and other minds.The TNGS has no difficulty in accounting for this lastprocess in terms of the connection between emotions,learning, and social influences on development of the self,at least in semantically equipped species.

Well before such social developments, the origin ofa sense of situatedness and of familiarity can be linkedto both the phenomenal and the voluntary motor aspectsof self-development. Of course, much remains to belearned about the detailed mechanisms underlying theperceptual categorization of states of the body. It is al-ready clear that memory systems, interacting with inter-


nal input from the various regulatory systems of the body,can render such categorizations omnipresent. Moreover,emotional responses interacting with value systems andthe homeostatic functions of the brain play key roles inboth primary and higher-order consciousness.

Finally, in considering the entries in Table 1, a fewadditional points are worth stressing. While the funda-mental and general properties of consciousness cannotbe relinquished, there are shifting degrees of contribu-tion from each of the informational and subjective prop-erties. These shifts are related to alterations in value sys-tems, learning experiences, emotions, and fluctuationsin attentional mechanisms. Obviously, the variationsacross conscious properties with experience can be verygreat depending on inputs to the dynamic core.

It is not difficult to imagine how, from such mix-tures and varying interactions among the properties inTable 1, one can discern the origins of complex mentalstates such as beliefs, desires, and emotional responses.Given experience and the existence of linguistic skills,it is perhaps not too far-fetched to imagine how evenlogical thought might have emerged from their interac-tions during experience. It remains to be seen howclosely such connections can be established. The obvi-ous point is that consciousness and its underlying C′states are central to all of these complex expressions,both rational and irrational.


IdentityT H E S E L F , M O R T A L I T Y , A N D V A L U E

n addition to providing an analysis of causationand the phenomenal transform, a theory of conscious-ness must account for subjectivity. This is not simpleidentity or individuality. It is the possession of a uniqueconscious history whose underlying neural states are ca-pable of refined discriminations that affect behavior atthe same time as they give rise to subjective feelings.

Given the nature of inheritance and evolutionaryselection, every multicellular organism may be said tohave a unique biological identity. In animals with anadaptive immune system, that identity is essential for

131

survival. But until cognitive systems arose in evolutionand consciousness appeared, the activity of a freely be-having self with richly idiosyncratic behavior, while im-pressive, was nonetheless limited. It is true that learningand communication systems arose in evolution well be-fore primary consciousness. Organisms such as bees orwasps can, for example, show remarkable adaptive be-havior in groups that depends to some degree on individ-ual variation. But the outcome in groups such as eusocialinsects is less autonomous and more statistical in naturethan the behavior of individual conscious animals.

We do not know at what point in evolutionaryhistory primary consciousness first arose. However, bycomparing homologous neural structures required forits expression in humans and other vertebrates (for ex-ample, a thalamocortical system and ascending valuesystems along with certain behavioral patterns), we canput forth a tenable conjecture that primary conscious-ness appeared in vertebrates first at the transition be-tween reptiles and birds, and second at the transitionbetween reptiles and mammals.

The ability to construct a scene related to thevalue-category history of an individual marks the ap-pearance of the self. An organism with a self can makerich discriminations based on its past learning historyand can use consciousness to plan, at least in the periodrepresented by the remembered present. The complexintegration of the dynamic core, modulated by behav-ioral history and the memories of individual learning

i d e n t i t y132

events, leads to adaptive behavior that is necessarily idio-syncratic for that individual organism.

The basis for individual emotional responses restswith the diffuse ascending-value systems, such as the lo-cus coeruleus, the raphe nuclei, the various cholinergicand dopaminergic systems, and several hypothalamicsystems. Other autonomic systems and brainstem nucleicontrolling bodily responses also provide a key basis forthe homeostatic, cardiorespiratory, and hormonal activi-ties that modulate emotion. In addition to the signalsprovided by these essential self-regulating systems of thebrain, there are also the proprioceptive and kinestheticresponses accompanying various movements. An indi-vidual with dawning primary consciousness already re-ceives “self-inputs” from such systems of motor control.As I already mentioned, it may even be that a spontane-ously moving fetus in late development distinguishes be-tween brain inputs arising from self-generated bodilymovements and those inputs generated by motions in-duced from without. There is enough evidence to makethe case that input from value systems and propriocep-tive systems can combine with modal sensory inputs toyield some of the earliest conscious experiences. It islikely that such fundamental adaptive systems remaincentral for the rest of an individual’s conscious life,whatever the additional qualia may be that develop withongoing experience.

If this is the case, the individual self necessarily hasa “point of view” that, given the activity of the dynamic

i d e n t i t y133

core, is integrated and generally persistent. Thus, if oneasks whether the appearance of a scene in primary con-sciousness has a “witness,” the answer is likely to be thatthe witness is constituted in an ongoing fashion by theintegrated bodily responses considered above as they re-late to those from memory and perceptual input. Indeed,it must be said that the idea of a “witness” is, to somedegree, anomalous: the first person is simply present.Given the continual sensorimotor signals arising from thebody, subjectivity is a baseline event that is never extin-guished in the normal life of conscious individuals. Butthere is no need for an inner observer or “central I”—inJames’s words, “the thoughts themselves are the thinker.”

Of course, in animals without semantic abilities,higher-order consciousness cannot be present. A self de-rived from primary consciousness is not able to symbol-ize its memory states, or become truly self-conscious orconscious of being conscious. When the necessary re-entrant circuits evolved in higher primates and finallyin Homo sapiens, a concept of the self appeared alongwith concepts of the past and future. Although as con-scious humans, we experience the Heraclitean illusionconsisting of a point in present time advancing from thepast into the future, a little reflection will show that anactual coupling to physical space-time can only occurin the “remembered present” of primary consciousness.For an animal with higher-order consciousness the pastand future are conceptual constructs.

One must avoid the temptation of splitting and

i d e n t i t y134

over-defining mental states and representations. Higher-order consciousness that is reflected by qualia in a high-dimensional space, and which is integrated to yield ascene having a focal center with shifting fringes and on-going changes, can never be focused to just one precisetoken. You cannot, for example, be aware of just a redspot and nothing else. The constructive integration thatleads to a unitary representation incorporating manydistinctions has more adaptive significance to the indi-vidual than any such limited designator or token, how-ever precise.

Thus, there is adaptive value in such multidimen-sional and situated discriminations. What they lack inabsolute precision, they make up for by enhancing ourability to generalize, to imagine, and to communicatein a rich environment. Higher-order consciousness maybe considered as a trade-off of absolute precision for richimaginative possibilities. Although our unitary con-scious scene is not necessarily veridical, for purposes ofplanning and creative scenario building it gains addedpower even as it gives up precision. I do not believe thatthis is an incidental point. The pervasive presence ofdegeneracy in biological systems is particularly notice-able in neural systems, and it exists to a high degree inthe reentrant selective circuits of the conscious brain. Incertain circumstances, natural languages gain as muchstrength from ambiguity as they do under other circum-stances through the power of logical definition. Associa-tion and metaphor are powerful accompaniments of

i d e n t i t y135

conscious experience even at very early stages, and theyflower with linguistic experience.

What is particularly striking about the operationsof the conscious human brain is the necessity for inte-gration, for a unitary picture, for construction, and forclosure. This is manifested by the obliviousness we haveto our blind spot, by various visual, somatosensory, andauditory illusions, and most strikingly, by neuropsycho-logical syndromes. The patient with anosognosia andhemineglect who denies ownership of a paralyzed lefthand and arm, the patient with somatoparaphrenia whoinsists that a touch on an anesthetic and paralyzed lefthand is a touch on her sister’s hand not on hers, or thepatient with alien hand syndrome—none of these indi-viduals is psychotic even if, in certain respects, each failsthe test of veridicality. The conscious brain in healthand disease will integrate what can be integrated andresists a fractured or shattered view of “reality.” I believethat these phenomena are reflections of the necessity forglobal reentrant circuits to form closed cycles with what-ever brain areas and maps are left to be integrated. Illu-sions can be induced in normal persons by manipulationof world and body signals. As I have already discussed,I believe that illusions reflect the shifts in patterns ofdominance among interactive reentrant maps. The take-home lesson is that our body, our brain, and our con-sciousness did not evolve to yield a scientific picture ofthe world. Instead, sufficient adaptation to an econicheis what saves the day, even in the presence of emotions

i d e n t i t y136

and imaginings that would be irrelevant or unavailableto a precise third-person description.

In animals with higher-order consciousness suchas ourselves, these operations provide a rich mixture ofimages, feelings, memories, pleasure and displeasure, be-liefs, and intentions—all of the intentional states as wellas those of moods. No two socially defined selves (neces-sarily socially defined in a speech community) will everhave identical brain states—the C′ states that entail Cstates. But such individuals can exchange informationeven on the basis of the mistaken belief that their Cstates are causal. This belief is safe, even if scientificallyfalse, for evolution has set up reentrant circuits to yieldC states as properties of C′ states. Indeed, C states arethe only reliable informational means we have of access-ing our C′ states.

This view is not paradoxical, nor does it qualify asdualism or as the weirdly detached epiphenomenologythat so distresses physicalist philosophers. C states arenecessarily entailed by C′ states, and the self has accessto the causal consequences of C′ states through C states.It will be a long time, if ever, before neurophysiologyusing on-line recording is so advanced that we can pre-dict the next C′ state (or C state, for that matter) inanywhere near a determinative fashion. Nonetheless,with further experimentation, larger patterns in C′ statesthat are neural correlates of consciousness will continueto be uncovered.

If this brain-based picture of how the self arises

i d e n t i t y137

turns out to be correct, there is, of course, one drearyconsequence: We are mortal. Once the substrate for Cstates is dissolved, the self, which is a dynamic process,ceases to be. There are some who find this conclusiondistasteful to the same degree as others find it repellantto accept the idea that we are not computers. Higher-order consciousness certainly allows for beliefs that arecontrary-to-fact. Let each self find its consolation in itsown way. Whatever belief system we espouse, the rich-ness of individual experiences during our lifetime con-tinues to be precious and irreplaceable.

There is a final point, which is not separate fromthe need for immortality or its dismissal. This concernsthe place of value in a world of facts. Scientific worldpictures based on the generality of physics alone haveno need for value, nor do they show evidence for it inthe inanimate universe. If our picture of conscious or-ganisms and of evolution is correct, however, value sys-tems are necessary constraints, both for evolutionaryselection and for neuronal group selection in animalswith higher brains. This does not mean, however, thathigher-order social values are genetically specified. In-stead, it means that such values will arise under the con-straints of adaptive systems, particularly those of con-scious individuals. While there is a biological basis forvalues, it is only through historical encounter and socialexchange as humans that we can build on such valuesto yield rights. On at least one planet in the universe,the evolutionary emergence of the reentrant dynamic

i d e n t i t y138

core with its C′ states has assured the place of value ina world of facts. Indeed, from a causal point of view,the reverse is also true—only as a result of value systemsin a selectional brain can the bases emerge for the phe-nomenal gift of consciousness.

i d e n t i t y139

Mind and BodyS O M E C O N S E Q U E N C E S

any confusions about the mind-bodyproblem are linguistic in origin. Others have to do withmisunderstanding of the procedures we must adapt instudying consciousness. Unlike physics, which assumesconsciousness and perception and takes a God’s-eyeview of its domain, the study of consciousness mustrecognize the first-person, or subjective, point of view.As a third-person observer studying another person’sconsciousness (see Figure 14), I must assume that thatperson has mental processes similar to my own. I must

140

then construct a variety of experimental procedures totest the reports made by the subject, searching for con-sistencies in his or her neural or psychological responses.

A theory of consciousness based on these effortsmust not conflict with the known laws of physics, chem-istry, or biology. Specifically, it must accept the fact thatthe physical world is causally closed—only forces andenergies can be causally effective. Consciousness is aproperty of neural processes and cannot itself act caus-ally in the world. As a process and an entailed property,consciousness arose during the evolution of complexneural networks with a specific kind of structure anddynamics. Before consciousness could emerge, certainneural arrangements must have evolved. These arrange-ments lead to reentrant interactions, and it is the dy-namics of reentrant networks that provide the causalbases that entail conscious properties. Such networkswere chosen during evolution because they provided an-imals with the ability to make high-level discrimina-tions, an ability that afforded adaptive advantages indealing with novelty and planning.

Consciousness reflects the ability to make distinc-tions or discriminations among huge sets of alternatives.These distinctions are made in fractions of a second andthey vary continually. As a series of phenomenal experi-ences, consciousness is necessarily private—it is tied toan individual’s body and brain and to the history of thatindividual’s environmental interactions. That history isunique—two different individuals of a species, even

m i n d a n d b o d y141

twins, cannot identically share the same conscious state.Indeed, the likelihood of any two conscious states beingidentical, even in one individual, is infinitesimally small.In this view, while no mental change can occur withoutan underlying neural change, the converse is not neces-sarily the case. Many neural changes have no effect onthe phenomenal character of the conscious state as re-flected in qualia.

Qualia are high-order distinctions, and the scenesof consciousness can be looked upon as a series of qualia.Such a series is experienced over a vast number of eventssignaling from the world, the body, and the brain itself.The multiplicity and potential parallelism of theseevents are both of high order, and the qualia that areincluded in an integrated but varying scene cover a widerange of experiences. They include perceptions, images,memories, sensations, emotions, moods, thoughts, be-liefs, desires, intentions, motor scenarios, and rich sig-nals—however vague—of bodily states. These variedexperiences may at first seem too disparate to be encom-passed by the mechanisms for the emergence of con-sciousness that are proposed here. But it must be re-membered that, in a highly connected complex systemlike the brain, integration of the combinatorial interac-tions of cortical and subcortical areas can result in anenormous number of states.

The function of the reentrant dynamic core,whether in primary or higher-order consciousness, canbe modulated by the brain mechanisms underlying focal


attention and memory. Subcortical structures such asthe thalamus and basal ganglia can mediate attentionalnarrowing of core states. In this sense, conscious statesdepend as heavily on nonconscious mechanisms of at-tention as they do on the nonconscious mechanisms ofperceptual categorization.

Inasmuch as consciousness arises as a result of re-entry in the dynamic core, it is necessarily integrated byreentry. To the subject, consciousness appears as a uni-tary process and, because of the binding and synchronythat result from reentry, the brain is constructive.

However, as I have mentioned, certain syndromesshowing aberrant conscious states can emerge from al-terations of the core and its interactions with noncon-scious substrates. In such pathological states, or even ininduced states such as those of a hypnotic trance, thecore can split into a small number of separate cores, oreven be remodeled constructively. Splitting of the coredefinitely occurs in disconnection syndromes that resultfrom cutting the corpus callosum and anterior commis-sure. It is also likely to be a main basis of dissociativesyndromes such as hysteria. Core remodeling can occurin neuropsychological syndromes such as blindsight,prosopagnosia, hemineglect, and anosognosia. In suchsyndromes, it is likely that the dominant reentrant re-actions of the core are redistributed constructively, re-sulting in a reallocation of conscious and nonconsciouscapacities.

In both normal and abnormal states, the brain of


an experienced individual attends continually to signalsfrom the body and the environment, but even more tosignals from itself. Whether in the dreams of REM sleep,or in imagery, or even in perceptual categorization, avariety of sensory, motor, and higher-order conceptualprocesses are constantly in play. Given the mechanismsunderlying memory and consciousness, both sensoryand motor elements are always engaged. For example,in perception, there are contributions of motor ele-ments—not acted out—that result from the premotorcontributions of global mappings. And in visual imag-ery, the same reentrant circuits used in direct perceptionare reengaged, but without the more precise constraintsof signals from without. In REM sleep, the brain trulyspeaks to itself in a special conscious state—one con-strained neither by outside sensory input nor by thetasks of motor output.

In all of these processes, primary consciousness iscontinually related to temporal change. It has a dia-chronic structure and is necessarily historical. Primaryconsciousness is, however, tied only to successive inter-vals of present time—the remembered present. The lagof up to five hundred milliseconds that is found betweenintended action, neural response, and conscious aware-ness is not a paradox if one understands the relationshipbetween nonconscious automaticity and conscious plan-ning. Consciousness is not involved in automatic motorprocesses (except during the learning leading to auto-maticity), but instead is related to planning and to the


creation of new combinations of already automaticroutines.

I have stated in the Preface of this small book thatmy hope is to disenthrall those who believe that the sub-ject of consciousness is exclusively metaphysical or nec-essarily mysterious. It is a Herculean task for conscious-ness studies to rid the stables of dualism, mysterianism,paranormal projections, and unnecessary appeals to asyet poorly characterized properties at different materialscales—for example, quantum gravity. Some but not allof this task relates to the use of language. In this account,for example, I must answer to the accusation that I havesubmitted to the paradoxes of epiphenomenalism. Thisnotion, a cousin of dualism and a prompting groundfor “zombie-speak,” must be reexamined. I believe thatthe difficulties with this notion have arisen becauseof the failure to attend to the neural correlates of con-scious properties. Inasmuch as the neural process C′ thatentails consciousness C is causal and reliable, we donot find ourselves faced with a paradox. C′ underliesthe ability to make distinctions in a complex domain,and C states, the properties entailed by C′, are thosedistinctions.

This relationship allows us to talk of C as if it iscausal. For most situations, this is not dangerous, giventhe reliability of the relationship. Only when we aretempted to abrogate physics or give to C mystical pow-ers is this procedure hazardous. The relation of en-tailment between C′ and C clarifies the issue and helps


define qualia as higher-order discriminations with dis-tinct and specific neural bases. A consciousness-freezombie, on these grounds, is logically impossible—if ithad C′ processes they would necessarily entail C. Ofcourse, I am aware of the fact that the clarification intro-duced by this analysis must be proven by actual experi-ments on the relation between C′ and C. But like theproportionality constant of mass in the equation F �

mA and the assumption of the constancy of the velocityof light in a vacuum, the foregoing analysis promisesa simplification and coordination of one of the mostchallenging problems of science.

Needless to say, I am aware of those who expectsuch a scientific analysis to explain the “actual feelingof a quale”—the warmness of warmth and the greennessof green. My reply remains the same: these are the prop-erties of the phenotype, and any phenotype that is con-scious experiences its own differential qualia becausethose qualia are the distinctions made. It suffices to ex-plain the bases of these distinctions—just as it sufficesin physics to give an account of matter and energy, notwhy there is something rather than nothing. This ourtheory can do by pointing out the differences in neuralstructures and dynamics underlying different modalitiesand brain functions.

Finally, some general remarks may be in order. Theview I have taken emphasizes the constructive, irrevers-ible, variable, yet creative properties of the brain. Theseare properties that can be explained on the basis of a


selectional theory of brain function such as neural Dar-winism. This theory argues against any simple-mindedreduction of historical events inasmuch as it is based onpopulation thinking and Darwinian evolution. More-over, it is worth pointing out that the occurrence of Cas an entailed property of C′ does not contravene es-thetic or ethical judgments inasmuch as the constraintsof conscious systems such as C′ depend ultimately onvalue systems.

In line with these reflections, I have previously sug-gested that there are two main modes of thought—logicand selectionism (or pattern recognition). Both are pow-erful, but it is pattern recognition that can lead to cre-ation, for example, in the choice of axioms in mathemat-ics. While logic can prove theorems when embedded incomputers, it cannot choose axioms. Even if it cannotcreate axioms, it is useful in taming the excesses of cre-ative pattern making. Because the brain can function bypattern recognition even prior to language, brain activitycan yield what might be called “pre-metaphorical” capa-bilities. The power of such analogical abilities, particu-larly when ultimately translated into language, rests inthe associativity that results from the degeneracy of neu-ral networks. The products of the ensuing metaphoricalabilities, while necessarily ambiguous, can be richlycreative. As I have stressed, logic can be used to tamethe excesses of those products, but cannot itself becreative to the same degree. If selectionism is the mis-tress of our thoughts, logic is their housekeeper. A bal-


ance between these two modes of thought and the end-less riches of their underlying neural substrates can besampled through conscious experience. Even if, some-day, we are able to embed both these modes in the con-struction of a conscious artifact and thus further extendour comprehension, the particular forms of conscious-ness that we possess as humans will not be reproducibleand will continue to be our greatest gift.


Glossary

Action potential. An electrical impulse traveling down aneuron’s membrane conducting signals from cell bodies tosynapses.

Adaptive immune system. The means by which verte-brates recognize foreign molecules, viruses, and bacteriaand react to them. Immune systems achieve this by con-structing a large repertoire of antibodies, each with a partic-ular potential binding site.

Anosognosia. A syndrome characterized by denial by a pa-tient of illness or unawareness of its existence. Seen particu-larly in some patients with strokes affecting the right corti-cal hemisphere.

Aphasia. Impairment or loss of the capacity to produce orunderstand language following brain damage.

Areas: prefrontal, parietal, temporal, visual, auditory, so-matosensory, motor. Regions of the cerebral cortex medi-ating one or more aspects of sensory or motor responses.

149

Areas are designated primary if they are the first to receiveprojections from the thalamus.

Association areas. Regions of cortex that do not includeprimary sensory or motor areas.

Associativity. The property of connecting or correlatingvarious brain regions, functions, or memories.

Attention. The ability to consciously select certain featuresfrom the vast array of sensory signals presented to thebrain.

Automaticity. The conversion of behavior after consciousrehearsal to nonconscious behavioral routines. Reflected insome aspects of procedural memory.

Autonomic nervous system. The visceral, mainly involun-tary, system of nerves consisting of sympathetic and para-sympathetic divisions controlling the internal environment.The first is for “fight or flight” reactions, the second is for“rest and digest.” Regulated mainly by the hypothalamus.

Axon. Structurally an extended neuronal process that car-ries action potentials to a synapse.

Basal ganglia. A linked collection of five large nuclei atthe center of the forebrain that serve to regulate motoracts and automatic responses that are nonconscious by in-teracting through the thalamus with the cortex.

Binding problem. How can different cortical areas andmodalities act synchronously and coherently (simulta-neously for movement, color, orientation, and so on) de-

g l o s s a r y150

spite the fact that each is specified by separate regions andthere is no superordinate or executive area? Plausiblysolved by reentry.

Binocular rivalry. The alternation over time of the percep-tion of two disparate inputs (for example, vertical bars versushorizontal bars) presented simultaneously to different eyes.

Bipedal posture. The upright stance in which the hindlimbs bear weight and carry out walking, freeing the upperlimbs from these functions.

Blindsight. In certain patients, conscious visual experienceis entirely lost for all or part of the visual field, yet the abil-ity to respond more or less accurately to visual stimuli un-der test conditions remains.

Brain dynamics. The functional (that is, electrical andchemical) complex of activities as distinguished from theanatomy within which these activities are carried out.

Brain scans. Various techniques for noninvasively follow-ing brain dynamics in living subjects. These techniques in-clude functional magnetic resonance imaging (fMRI) andmagnetoencephalography (MEG).

Brainstem. A part of the brain made up of the thalamus,the hypothalamus, the midbrain, and the hindbrain. Thehindbrain includes the cerebellum, pons, and medulla, butthe cerebellum is generally excluded from the definition.

Broca’s area. An area in the left frontal lobe, damage to

g l o s s a r y151

which leads to difficulties in speech production or motoraphasia.

C and C′. C designates any conscious process; C′ desig-nates its underlying neural activity.

Causal efficacy. The action in the physical world of forcesor energies that lead to effects or physical outcomes.

Causality. According to the laws of physics, the causal or-der is closed—that is, it cannot be affected directly bymental properties such as qualia.

Cell body. The part of a neuron containing the nucleuswith its DNA.

Cell migration. The patterned movement of neurons ortheir cellular precursors during the formation of the brain.

Cerebellum. A large structure attached to the brainstemthat contributes to the coordination of motor activity.

Cerebral cortex. A six-layered mantle of neurons (graymatter) over the surface of the cerebral hemispheres. Themantle is folded into convoluted protrusions (gyri) andclefts (sulci).

Channel. A molecular structure (a protein) in the cellmembrane that allows ions to pass from one side to theother.

Cholinergic nuclei. Collections of neurons activated bythe neurotransmitter acetylcholine.

Clone. The asexual progeny of a single cell.

g l o s s a r y152

Closure; filling-in. The tendency of the brain to integratesignals with whatever interactions are available to it. Filling-in is found in the failure to notice the blind spot; otherexamples include cases of denial such as those found inanosognosia.

Coarticulated sounds. Sounds that simultaneously havecomplex mixtures of different frequencies and energies,such as human speech or vocal sounds.

Coherence. The simultaneous or synchronous activity ofdistant collections of neurons or other agents.

Combinatorial. Mathematical operations quantitatively de-scribing the various interactions of different systems orparts.

Complexity. A property of any system composed of multi-ple heterogeneous smaller parts that nonetheless interact togive integrated outcomes.

Computations. In a narrow sense, the actions carried outby a digital computer processing algorithms via a program.

Computer. A device (here, a digital computer that consistsof hardware and software) utilizing collections of algo-rithms or effective procedures making up a program thatperforms logical operations to yield an output. See also Tu-ring machine.

Concatenated reentrant loops. Reentrant structures thatform cycles that overlap with one another.

Concept. Usually refers to propositions expressing abstract

g l o s s a r y153

or general ideas. Used here to refer to the brain’s ability tocategorize its own perceptual activities and construct a“universal.”

Corpus callosum. A large fiber system connecting similarareas of the right and left cerebral hemispheres. Cutting ordestroying this bundle leads to the disconnection syn-dromes shown by split-brain patients.

Correlation. A statistical term used to describe and quan-tify nonrandom relations between two systems.

Cortex: primary, secondary, tertiary. Somewhat old-fashioned terms distinguishing the portions or areas of thecortex receiving direct sensory input or mediating directoutput (primary cortex) from “higher” areas that connectto these areas (association cortex).

Cortical hemispheres. The two (left and right) large struc-tures, making up a good portion of the forebrain, whichhave the cerebral cortex as their mantle.

Corticostriatal. Axonal projection from the cortex to theinput nuclei of the basal ganglia (the so-called striatum).

Covariance. A statistical term quantifying the mutualchange in two or more variables.

Degeneracy. The ability of different structures to carryout the same function or yield the same output.

Dendrite. One of the many postsynaptic (input) branchesof a neuron that receives axonal connections to form a syn-apse at sites called dendritic spines.

g l o s s a r y154

Developmental selection. The first tenet of the TNGS. Itrefers to the creation of large repertoires of variant circuitsin the microanatomy of the brain during development.

Differentiable; differentiated. Used to refer to the factthat conscious experience can change from one unitaryscene to another in an apparently limitless fashion. SeeUnitary.

Direct and indirect pathways. Two main routes of con-nections within the basal ganglia leading to stimulation orinhibition of thalamic nuclei by basal ganglion activity.

Discrimination. The capacity of conscious systems to cate-gorize, distinguish, or differentiate among a multitude ofsignals or patterns in terms of integrated scenes and qualia.

Distributed system. Widespread and separated neuronalgroups that can nonetheless interact through connectionsto give an integrated response or output (see Complexity).The cortex is a distributed system.

Dopaminergic nuclei. Four major systems of brain neu-rons utilizing the neurotransmitter dopamine. They consti-tute a value system involved in the reward systems of learn-ing. Dopaminergic synapses are considered targets ofantipsychotic drugs, particularly for schizophrenia.

Dualism. The belief that the facts of the world cannot beexplained except through the existence of two differentand irreducible principles. In philosophy, its most notedexponent was Rene Descartes, who divided the world into

g l o s s a r y155

res extensa (subject to physics) and res cogitans (not so ac-cessible).

Dynamic core. A term used in the extended TNGS to re-fer to a system of interactions, figured mainly in the thala-mocortical system, which behaves as a functional cluster.The core sends signals mainly to itself, and its reentrant in-teractions are assumed to give rise to conscious states.

Econiche. The part of the environment in which a speciesacts and within which natural selection occurs.

Embodiment. The view that the mind, brain, body, andenvironment all interact to yield behavior. Used in somesense to contradict the idea of a “disembodied mind” ordualistic consciousness.

Emotion. The complex of feeling, cognition, and bodilyresponses reflecting the action of value systems within theconscious brain. The term comprises a huge range of re-sponses, of which the examples are obvious and known tothe reader.

Enkephalin. A small peptide, member of a set of endoge-nous opioids (opiate-like substances produced by thebrain). Their actions can produce analgesia or blunting ofpain.

Entailment. A relationship of implication used here in re-gard to the relationship between C′ processes and C. C′ en-tails C as a property. Thus, C (consciousness) is entailedby physical processes in the brain, mainly in the dynamiccore.

g l o s s a r y156

Entropy. In physics, a measure of order or disorder. In in-formation theory, a measure of the reduction of uncer-tainty. Entropy can be related to the number of differentstates of a system weighted by their probability of occur-rence.

Epigenetic. Biological processes that do not depend di-rectly upon gene expression. An example: “neurons thatfire together, wire together.”

Epiphenomenal. Lacking in causal effectiveness. For exam-ple, the lights flashing on a computer console could be re-moved without affecting its internal processing. Philoso-phers argue about whether C is epiphenomenal and somefind this paradoxical. This book makes the case that, prop-erly understood, there is no paradox.

Episodic memory. The memory of past events, mediatedby the interaction of the hippocampus with the cerebralcortex. Removal of the hippocampus obliterates the capac-ity to form such memories from the time of the operation(or lesion) forward.

Evolution. The process underlying the emergence and sur-vival of living things. Accounted for by several theories at-tributable to Charles Darwin, of which the central one isnatural selection.

Experiential selection. The second tenet of the TNGS,which states that a secondary repertoire of functioning neu-ral circuits is formed on the basis of existing neuroanat-

g l o s s a r y157

omy by means of the selective strengthening and weaken-ing of synaptic efficacies.

Explanatory gap. A term used by philosophers to stressthe difficulty or impossibility of relating conscious phenom-ena to the neural workings of the brain.

Feedback. A term used strictly in control theory to desig-nate correction of a deviation in output by an error signalderived from a sample of that output. For example, if anamplifier is to amplify a sine wave and the output is dis-torted, an error signal is sent back through a single chan-nel to modify the dynamics to yield the correct waveform.Feedback always operates from an output to an earlierstage, whereas reentry may occur between stages operatingin parallel at the same or different levels in a system. Inpopular use, the term feedback is applied indiscriminatelyand often vaguely to any correction of input by a reversesignal.

First-person experience. The privacy of an individual’sstream of consciousness, which cannot be shared directlyby a third-person observer.

fMRI (functional magnetic resonance imaging). A non-invasive scanning method for observing brain dynamics byusing magnetic resonance imaging to record changes inblood oxygen levels that are correlated with neural activity.

Focal attention. An attentional state directed strongly to-ward a single object, thought, or experience.

Fourier transform. A mathematical operation for dealing

g l o s s a r y158

analytically with functions (waves, for example) by con-verting them to sums of sine and cosine functions.

Frequency tag. A method used in MEG and EEG (electro-encephalography) to mark a brain response as one re-flecting a given signal. In MEG, for example, oscillating asignal at 7 Hz (seven times a second) will show a sharpspike at that frequency in an ensuing record of brain activ-ity analyzed by Fourier transforms.

Freudian unconscious. A domain of which a subject isnot conscious but that is capable of being made consciousby psychoanalytic techniques.

Fringe. Used by William James to denote the “influenceof a faint brain process on our thoughts,” as it makes us“aware of relations and objects but dimly perceived.”

Functional cluster. In complexity theory, a system or partof a system that interacts mainly with itself. The dynamiccore is a functional cluster.

Functional connectivity. Paths within a neuroanatomicalnetwork that actually reflect neural dynamics, for example,input yielding output.

Functional segregation. Relative restriction of the activityof a brain area to a given function. For instance, one vi-sual area may be functionally segregated for colors, anotherfor object movement, and so on.

GABA. Gamma-aminobutyric acid. An inhibitory neuro-

g l o s s a r y159

transmitter found, for example, in the local inhibitory cir-cuits of the cortex and those of the basal ganglia.

Genetic code. The set of rules by which DNA sequencesspecify amino-acid sequences in proteins. The code con-sists of non-overlapping triplets (for example, AUG for me-thionine and UUU for phenylalanine). There are sixty-four triplets, or codons, and twenty amino acids, and thecode is therefore degenerate (see Degeneracy).

Gestalt phenomena. Aspects of perception in which sim-ple sensory inputs can be grouped in particular ways to cre-ate a gestalt, a figure or form that is not a property of theobserved object but reflects the constructive capabilities ofthe brain.

Gestural communication. Exchanging messages throughgestures, as in mime or, when syntactically organized, signlanguage.

Glia. The supporting cells of the nervous system necessaryfor biochemical and energetic as well as structural func-tions, but not for signaling by action potentials. There areseveral types, including astrocytes and oligodendrocytes.

Global mapping. A term referring to the smallest struc-tures in the brain capable of perceptual categorization.Reflects the activity of multiple reentrant maps, motor andsensory, linked to nonreentrant structures and finally tomuscles and sensory receptors capable, through movement,of sampling a world of stimuli.

Globus pallidus. A part of the basal ganglia. It receives

g l o s s a r y160

connections from the caudate nucleus and sends projec-tions to the ventrolateral nucleus of the thalamus.

Glutamate. The main excitatory neurotransmitter of thecentral nervous system.

Hebb synapse. Named after Donald Hebb, a psychologistwho enunciated Hebb’s rule: When an axon of cell A ex-cites cell B and persistently takes part in firing it, a changeoccurs in one or both cells so that A’s efficiency in firingB is increased. A Hebb synapse follows this rule.

Hemineglect. Certain patients with damage to the rightparietal cortex no longer pay attention to or seem aware ofthe left side of a scene.

Heraclitean illusion. The notion of a point in timesmoothly flowing from the past through the present to thefuture. This is illusory in that only the present is directlyaccessible to experience, whereas the past and the futureare concepts.

High-dimensional space. We live in four-dimensionalspace-time (three of space, one of time). Qualia space is ahigh-dimensional or n-dimensional space, where n is thenumber of axes along which distinctions can be made; n ismuch greater than four.

Higher-order consciousness. The capability to be con-scious of being conscious. This capacity is present in ani-mals with semantic abilities (chimpanzees) or linguisticabilities (humans), and those with linguistic abilities arealso able to have a social concept of the self and concepts

g l o s s a r y161

of the past and future. Distinguished from primary con-sciousness.

Hippocampus. A sausage-shaped neural structure lyingalong the anteromedial region of each temporal lobe. Incross section it resembles a sea horse, hence the name. Nec-essary for episodic memory.

Homeostatic. Tending to maintain constancy in the statusof the interior environment, whether of cells or tissues.

Hominines. A group within the order Primates that in-cludes modern humans and their antecedents, who ap-peared after the divergence from precursors of the apes.

Homologous structures. Distinct structures evolutionarilyderived from a common ancestor, in respect to structureor function. The thalami of dogs and mice are homolo-gous to those of humans.

Hox and Pax Genes. Ancient genes that regulate morpho-genesis. Pax 6, for example, is essential for normal develop-ment of the eye. Hox genes regulate the structures of thehindbrain. Their expression in the embryo is place depen-dent.

Huntington’s disease. A hereditary disease involving thedegeneration of the caudate and putamen of the basal gan-glia. It leads to continual involuntary movements (chorea)and progressive dementia, and ends with death.

Hypothalamus. A set of nuclei below the thalamus that af-

g l o s s a r y162

fect feeding, sex, sleep, emotional expression, endocrinefunctions, and even movement. A value system.

Ideal or perfect gas. A theoretical construct consisting ofrandomly colliding particles whose collisions are perfectlyelastic and exchange no mutual information.

Identity. All animals that are not twins are geneticallynon-identical, and each individual is therefore unique.This can be the case without having a conscious self.

Illusion. Psychophysically manipulated signals that lead toa perception of features not physically verifiable. They are“false” expressions of “real” sensory input. The illusory con-tours of a Kanizsa triangle are an example, and so is aNecker cube. Illusions can occur in various modalities andrange from simple to complex.

Information. In present usage, the reduction of uncer-tainty conveyed by a message.

Inhibitory loops. Neurons linked by inhibitory synapsesand connected in loops. The classical example is providedby the basal ganglia, whose polysynaptic loops either canbe inhibitory or can inhibit inhibition (disinhibition).

Instructionism. The idea that information from structuresthat are to be recognized is necessary in constructing a rec-ognition system. An example is the idea, now disproven,that antibodies are specific because they wrap around theshapes of antigens when these antibodies are formed. Theopposite of instructionism is selectionism as represented bythe theory of evolution and the TNGS.

g l o s s a r y163

Integration. In complexity theory, the measure of mutualinformation or entropy reduction in a system. In brain sci-ence, the relating, correlation, or connection of signals toyield a unitary output.

Intensity. A measure of strength. In electromagnetic mea-surements, such as magnetoencephalography, the inten-sity is the square root of the power.

Intentionality. The idea proposed by Franz Brentano thatconsciousness refers to particular objects—it is aboutthings. This is not the same as “intending.”

Intralaminar nuclei. Nuclei of the thalamus that projectdiffusely to the frontal cortex, caudate, and putamen. Prob-ably concerned with setting thresholds of their targets andthus implicated in the maintenance of consciousness.

Irreducibility. A theory or statement is irreducible if itcannot be fully accounted for by a theory at some lowerlevel of organization.

Jamesian properties. Consciousness is a form of aware-ness, is continuous but continually changing, is private,has intentionality, and does not exhaust the properties ofits objects.

Kinesthetic. Relating to perception of movement or posi-tion of joints, or limbs, or body.

Language. In strict use, a vehicle of communication thathas phonology (or signing), semantics, and syntax. Hu-

g l o s s a r y164

mans are the only species with true language. See Higher-order consciousness.

Lateral geniculate nucleus. A specific thalamic nucleusthat receives input from the optic nerve and sends projec-tions to the primary visual cortical area, V1.

Lexicon. A collection of tokens or words in the memoryof a semantically equipped or linguistic animal.

Linguistics. The study of language—that is, of phonol-ogy, semantics, and syntax. Neurolinguistics examines thebrain bases of true language.

Locus coeruleus. The slightly blue neuronal nuclei in themidbrain whose diffuse ascending projections to the thala-mus and cortex release noradrenaline. It is a value systemimportant in detection of salient signals and in sleep.

Logical atomism. The notion proposed by Bertrand Rus-sell and Ludwig Wittgenstein that minds can be con-structed out of sensations and images and that, by con-structing everything from simpler entities, we will have acomplete description of what is the case. Wittgenstein deci-sively rejected this notion in his later life.

Long-term memory. Memory system with durationslonger than working or short-term memory. Episodicmemory is an example.

Machine. A device constructed of parts to perform a de-fined function. The most general example, perhaps, is aTuring machine.

g l o s s a r y165

Magnetoencephalography (MEG). The use of supercon-ducting quantum interference devices (SQUIDS) to mea-sure minute magnetic fields in the living brain. The de-vices used have multiple electrodes covering the entirebrain and are very sensitive to temporal changes in inter-nal currents arising from as few as twenty thousand neu-rons. Resolution in space is on the order of 1–1.5 centime-ters, as compared with fMRI, which can get down to 3–4millimeters or less but lacks the temporal resolution ofMEG.

Maps. Maps in the brain are either topographic or nontop-ographic, meaning that they either do or do not conservethe geometrical relationships of their neighboring parts. Inthe first case, the term refers to projections from severalcells in a neighborhood to another neighborhood—pointto area or area to area. A key example is provided by theretinal maps in the thalamus, which map in turn to corti-cal area V1. Reentry among maps having different func-tions binds them into dynamic integrated structures.

Meaning. In neurobiology, the realization of a value sys-tem’s bias or of a goal. In language, the denotation andconnotation of a word—its semantics.

Memory. Term used for a variety of systems in the brainwith different characteristics. In all cases, however, it im-plies the ability to reinvoke or repeat a specific mental im-age or a physical act. It is a system property that dependson changes in synaptic strengths.

g l o s s a r y166

Mental images. The creation by the brain of images, with-out external stimulation by original objects. Mental rota-tion is the ability consciously to turn a mental image to anew orientation.

Mental representations. A term used by some cognitivepsychologists who have a computational view of the mind.The term is applied to precise symbolic constructs orcodes corresponding to objects and, by their computation,putatively explaining behavior.

Metaphor. Figure of speech in which a term is used todesignate an object it does not ordinarily refer to; “the eve-ning of life” is a dictionary example. The brain origin ofmetaphorical reference may relate to embodiment.

Metastable. Stable in the complete absence of perturba-tion, a state typically realizable for only a short period oftime but with definite structure while it lasts.

Millisecond. One thousandth of a second. Synapses func-tion in the range of one to ten milliseconds.

Mime. Gestural communication without syntax or arbi-trary symbolism.

Mind. The totality of all conscious and underlying uncon-scious processes originating in the brain and directing allbehavior. In the philosophical sense, it is one part of theseesaw known as the mind-body problem—How is it pos-sible for brain activity to give rise to mental activity?

Modularity. The doctrine that the brain functions largely

g l o s s a r y167

by having different regions or modules that perform dis-tinct functions. This view results in “localizationism,” asopposed to the contrary view of “holism”—that the wholebrain is required. Both views disappear when considered interms of selectionism and complexity theory.

Modulation. Adjustment, adaptation, and regulation caneach give rise to modulation. In electronics, variation ofthe amplitude, frequency, or phase of a signal.

Motor regions. A variety of cortical areas—including theprimary motor cortex, premotor areas, frontal eye fields—that can give rise to muscle contractions when simulated.

Mutual information. In statistical information theory, themutual change in entropy upon interaction of any twoparts of a system.

Natural selection. Chief among the theories of evolutionformulated by Darwin. Effectively, the idea that competi-tion among variants in a population leads to differential re-production, with concomitant changes in gene frequency.

Neural correlate of consciousness. Nerve activity that isfunctionally correlated with conscious states.

Neural Darwinism. A term applied to the TNGS to em-phasize the application of selectionism and populationthinking to the brain.

Neuromodulator. One of the substances that alter synap-tic action, including a variety of neurally active peptidesthat can cause inhibition or excitation when applied to tar-

g l o s s a r y168

get neurons. There is a very large number of these sub-stances, and they can affect pain, emotions, endocrine re-sponses, and responses to stress.

Neuron. A nerve cell of the central or the peripheral ner-vous system.

Neuronal group. A tightly interactive local collection ofneurons (hundreds to thousands), both excitatory and in-hibitory. It is the unit of selection in the TNGS.

Neurophysiology. The study of the detailed electrical(and associated chemical) responses of neurons, singly orin whole systems. The experiments in this field range fromtissue culture of neuronal cells to slices of the brain toprobes with electrodes of whole neural areas in behavinganimals.

Neurotransmitters. The chemicals released into the synap-tic cleft from vesicles in the presynaptic neuron that thenbind to receptors in the postsynaptic neuron, changing itstransmembrane electrical potential or intracellular chemis-try. Neurotransmitters are the chief means of communica-tion between neurons. See GABA and Glutamate.

Noise. Used in electronics and information theory to referto random or uncorrelated perturbations on a signal.

Nonconscious. Refers to brain activities unable to becomeconscious, in contradistinction to the Freudian uncon-scious.

g l o s s a r y169

Nonself. Refers to all signals not transmitted directly bythe body: signals from the environment.

Nuclei. Closely connected collections of neurons with simi-lar activities, functions, neurotransmitters, and input-output relations that have a definite neuroanatomicalboundary.

Optic nerve. The main set of fibers coming from the gan-glion cells of the retina and projecting to the lateral genic-ulate nucleus.

Output nuclei of the basal ganglia. The internal segmentof the globus pallidus and the pars reticulata of the sub-stantia nigra, projecting to the thalamus.

Parkinson’s disease. A motor-system disease resultingfrom a loss of dopaminergic neurons of the substantia ni-gra. It is characterized by tremors, muscular rigidity, al-tered gait, and also occasionally by cognitive impairment.

Perceptual categorization. The process by which thebrain “carves the world up” to yield adaptive categories.The most fundamental of early cognitive functions.

Perfect crystal. A crystal without any irregularities in itsinternal order. The third law of thermodynamics statesthat the entropy of a perfect crystal of a pure substance atzero degrees absolute is zero.

Phenomenal experience. The experience of qualia; con-sciousness.

Phenomenal transform. A term used here to designate

g l o s s a r y170

the process by which neural activity in the reentrant dy-namic core (C′) entails the phenomenal property of con-sciousness (C).

Phonetics. The study of the sounds of speech. It is in-cluded in phonology, which also includes phonemics, thestudy of the smallest units of speech.

Phrenology. A discredited system of assignment of higherfaculties to modules or particular regions of the brain diag-nosable by bumps on the head. Founded by Joseph Gall.

Piagetian notion of self. After Jean Piaget, a noted devel-opmental psychologist. Applied here to the stage at whicha child can differentiate his or her own motor acts frommotions imposed from without.

Population thinking. Darwin’s seminal notion that spe-cies arise “from the bottom up” by selection of variant in-dividuals in a population.

Postsynaptic neuron. Neuron whose properties arechanged after release of neurotransmitter by presynapticneurons.

Power. Refers to the calculated energy distribution afterFourier analysis of waveforms such as those studied bymagnetoencephalography. It is the square of the intensity.

Premotor regions. Areas of the cortex that prepare themotor systems for movement. Another such area is the sup-plementary motor area, which aids in programming se-quences of motor acts.

g l o s s a r y171

Presynaptic neuron. The neuron that releases neurotrans-mitters into synaptic clefts after an action potential arrivesat the synapse.

Primary consciousness. The fundamental consciousness,which is proposed to arise first from reentry between re-gions carrying out perceptual categorization and those me-diating value-category memory. Results in the creation of ascene in the remembered present. See Higher-order con-sciousness.

Privacy. The fact that consciousness is experienced as afirst-person event that is not fully capable of being shared.

Procedural memory. A form of memory concerned withsequences of action or particular movements, such as bicy-cle riding. It is separate from episodic and semanticmemory.

Process. A series of changes. William James emphasizedthat consciousness is a process, not a thing.

Progenitor cells. Cells in the brain capable of giving riseto neurons. They are seen in olfactory regions and in thehippocampal region of adults.

Propositional attitudes. A philosopher’s term for beliefs,desires, and intentions.

Proprioceptive. Providing information about the relativepositions of the body in space and the relation of the bodysegments to each other.

Prosopagnosia. The inability to recognize a face, even a

g l o s s a r y172

previously familiar face, following cerebral damage. Recog-nition of other objects is not necessarily affected.

Protosyntax. Movement sequences and basal ganglion re-sponses that have ordered structures like syntax.

Putamen. A nucleus of the basal ganglia.

Quale; qualia. Terms used to refer to the “feel” of con-sciousness experience—“what it is like to be x,” where x,for example, is a human or a bat. I use the term “qualia”as coextensive with conscious experience. Consciousness re-flects the integration among vast numbers of qualia. Qua-lia are discriminations made possible by the activity of thereentrant dynamic core.

Qualia space. A construct reflecting the fact that qualiacannot be completely isolated but exist together in a multi-dimensional or high-dimensional space.

Raphe nucleus. A collection of cell groups in the midlineof the brainstem projecting to forebrain structures and re-leasing serotonin. A value system.

Recategorical. The process by which memory as a systemproperty interprets current input based on past experi-ence—that is, it does not replicate an original experienceexactly.

Receptors. Proteins on the surfaces of cells that bind vari-ous chemical ligands including, for example, neurotransmit-ters, neuromodulators, hormones, and drugs.

Reciprocal fibers. Axonal bundles connecting two regions

g l o s s a r y173

of the brain in both directions. These provide the anatomi-cal basis for reentry.

Reentry. The dynamic ongoing process of recursive signal-ing across massively parallel reciprocal fibers connectingmaps. This process results in binding and is the basis forthe emergence of consciousness through the workings ofthe dynamic core. Allows coherent and synchronous eventsto emerge in the brain; that is, it is the basis for spatiotem-poral correlation.

Reflex. Automatic sensorimotor loops; these are seenclearly, for example, in motor responses mediated by thespinal cord. They are nonconscious, and can be developedin the higher brain by conditioning.

Remembered present. A phrase used to describe the tem-poral aspect of the scene constructed in primary conscious-ness, suggesting the role of memory processes in that con-struction. Close to the specious present quoted byWilliam James in The Principles of Psychology.

Repertoire. A set of variants in a selectional system.

Representations. Results of conscious discriminations andclassifications; does not imply that the underlying neuralstates are representations.

Res cogitans. Descartes’ term for “thinking substance,” in-accessible to physical investigation. One component of sub-stance dualism.

g l o s s a r y174

Res extensa. Extended things—the other end of substancedualism—one accessible to physics.

Reticular nucleus. A structure surrounding the thalamusthat is also part of it, consisting mainly of inhibitory con-nections to specific thalamic nuclei.

Retina. The thin layer of photoreceptor cells, rods, andcones as well as ganglion cells in the eye that direct signalsto the optic nerve. Along with olfactory epithelia, this isthe part of the brain closest to the surface of the body.

Scene. The integration of inputs in a discriminatory fash-ion in primary consciousness.

Schizophrenia. Psychotic disease showing deep disordersof cognitive function, confusion, and splitting of thoughtand emotions. Not yet proven to be due to a specific de-fect of brain function but certainly a disease of conscious-ness.

Selectionism. The notion that biological systems operateby selection from populations of variants under a varietyof constraints. The opposite of instructionism.

Self. A term used to refer to the genetic and immunologi-cal identity of an individual, but more to the point of thisbook it refers to characteristic inputs from an individualbody related to its history and value systems. In its mostdeveloped form, seen with higher-order consciousness, itis a social self related to interactions within a speech com-munity.

g l o s s a r y175

Semantic memory. Memory related to identification of ob-jects, persons, places, and circumstances. Not episodic,however.

Semantics. The linguistic study of meaning and reference.

Sensorimotor loops. Connections between input signalsand motor activities such as are seen in global mapping.

Sensory receptors. Specialized neurons for various modal-ities such as sight (rods, cones), hearing (hair cells), smell(odorant receptors), and so on.

Short-term memory. An example is the memory for tele-phone numbers, generally considered to be limited toseven digits, plus or minus two.

Situatedness. Presence in an environment or econiche andawareness of it.

Sleep; REM sleep. The change of state characterized bydistinct changes in EEG, isolation of the brain from exter-nal input, and blockade of motor output. In rapid eyemovement (REM) sleep, there is an EEG pattern of lowamplitude, aperiodic fast spikes, and also dreaming. Aform of consciousness occurs in REM sleep.

Somatoparaphrenia. Failure to correctly identify bodyparts as one’s own.

Spatiotemporal correlation. According to the TNGS, inthe absence of logic such as governs a computer, the brainmust correlate time and space and sequence. It does thisthrough reentry.

g l o s s a r y176

Specific thalamic nuclei. Nuclei of the thalamus receivingsensory signals for different modalities (see Lateral genicu-late nucleus, for example) or signals for motor controlsuch as those from the basal ganglia. Specific nuclei do notconnect to each other but project to the cortex.

Specious present. The term quoted by William James inThe Principles of Psychology and used to designate the pres-ent that we are aware of experiencing. See Rememberedpresent.

Speech community. A group of individuals communicat-ing over time via a particular language.

Stochastic. Subject to random processes or noise.

Striatum. The input region of the basal ganglia consistingof the caudate nucleus and the putamen.

Subcortical. Refers to structures lying below the neocor-tex, such as the basal ganglia, the hippocampus, and thecerebellum, among others.

Subjectivity. Refers to the private self, and collectively thefirst-person experiences of such a self.

Substance P. A neuromodulator that can activate pain re-ceptors.

Substantia nigra. One of the basal ganglion nuclei con-taining cells expressing the neurotransmitter dopamine.

Subthalamic nucleus. Part of the basal ganglia. Lesions ofthis nucleus cause uncontrollable movements called bal-lismus.

g l o s s a r y177

Supervenient. A philosophical term to describe the rela-tion between C′ and C meaning roughly “dependent on.”A change in mental state therefore would necessarily re-quire a change in neural state.

Supralaryngeal space. The space in the throat resultingfrom developmental descent of the larynx in humans andallowing a great expansion and refinement of speechsounds.

Synapse. The critical connecting structure between neu-rons, which mediates their signaling by electrochemicalmeans (see Neurotransmitters; Postsynaptic neuron; Pre-synaptic neuron).

Synaptic strength. The degree by which neurotransmitterrelease affects postsynaptic response. Synaptic strengthmodification is a change in a synapse that weakens it orstrengthens it, altering communication between neuronsnecessary for the establishment of memory. Changes ofthis kind reflect neural plasticity.

Synaptic vesicles. Membraneous structures containing neu-rotransmitters at the axonal terminals of presynaptic neu-rons.

Synchrony. Simultaneity of events, such as simultaneousfiring among neurons.

Syntax. The study of grammar and ordering in linguistics.

Thalamus. The chief sensorimotor relay structure to thecortex. It is a key part of the thalamocortical system and

g l o s s a r y178

the dynamic core. Consists of specific nuclei, intralami-nar nuclei, and the reticular nucleus.

Third-person experience. The position of an external ob-server unable to directly experience another’s first-personsubjectivity.

TNGS. The theory of neuronal group selection, whichconsists of three tenets: (1) Developmental selection and(2) Experiential selection, both of which operate on reper-toires of neural variants, and (3) Reentry, a key processassuring spatiotemporal correlation and conscious integra-tion. It is a global brain theory explaining diversity and in-tegration in the central nervous system.

Token. A semantic element or word in a lexicon.

Turing machine. A finite-state automaton capable of read-ing, writing, and erasing zeros and ones from an endlesstape and then moving one space to the right or left undercontrol of a program. Turing machines are theoretical con-structs, and a theorem by Alan Turing showed that a uni-versal Turing machine could carry out any computationbased on effective procedures or algorithms.

Unconscious. The state of being unaware. See also Non-conscious and Freudian unconscious.

Unitary. The all-of-a-piece nature of a conscious scene,which cannot voluntarily be broken up into separate parts.

V1, V2, V3, V4, V5. The various areas of the striate andextrastriate regions of the cortex subserving vision.

g l o s s a r y179

Value; value systems. The constraining elements in a se-lectional system consisting in the brain of diffuse as-cending systems such as the dopaminergic system, the cho-linergic system, and the noradrenergic system of the locuscoeruleus. Value systems also include the hypothalamus,the reticular activation system, and the nuclei around theperiaquaductal gray matter of the brainstem. In humans,value is modifiable under some constraints.

Value-category memory. According to the extendedTNGS, this memory system involves fast synaptic changeleading to categories and it is altered by modulation origi-nating in value systems. Reentrant interactions of value-category memory with systems of perceptual categorizationlead to primary consciousness.

Variability. Refers to changes in brain responses at all lev-els, providing a basis for neuronal group selection.

Veridical. Matching physical reality as tested by scientifictheory and measurement.

Wernicke’s area. The posterior part of the superior tempo-ral gyrus (area twenty-two), which, when damaged, canlead to the inability to produce meaningful speech or com-prehend it—a condition called Wernicke’s aphasia. Seealso Broca’s area.

Zombie. A hypothetical humanlike creature that lacks con-sciousness but which, it is erroneously assumed, can carryout all of the functions of a conscious human.

g l o s s a r y180

Bibliographic Note

As I mentioned in the Preface, I have deliberatelyavoided specific references within the body of the text. A brieflist of references to orient the reader may nonetheless be use-ful here.

For descriptive insight, nothing beats the efforts of Wil-liam James:

James, W. The Principles of Psychology. Cambridge: HarvardUniversity Press, 1981.

James, W. “Does Consciousness Exist?” In Writings of Wil-liam James, edited by J. J. McDermott. Chicago: Univer-sity of Chicago Press, 1977, pp. 169–183.

An excellent perspective from a more modern stand-point may be found in:

Searle, J. R. The Mystery of Consciousness. New York: NewYork Review of Books, 1997.

This publication collects Searle’s reviews of various booksin the field and at the same time nicely captures the salient

181

issues. The issue of privacy is captured well by another philos-opher:

Nagel, T. Mortal Questions. New York: Cambridge Univer-sity Press, 1979.

For a different approach to similar issues, see:

Kim, J. Mind in a Physical World. Cambridge: MIT Press,1998.

For a psychological theory concerned with consciousaccess, see:

Baars, B. J. A Cognitive Theory of Consciousness. Cambridge,England: Cambridge University Press, 1988.

My own efforts to build a scientific theory range overmore than two decades. They are described in a series ofbooks and papers that contain extensive bibliographies com-posed with scholarly attributions in mind:

Edelman, G. M., and Mountcastle, V. B. The Mindful Brain:Cortical Organization and the Group-Selective Theory ofHigher Brain Function. Cambridge: MIT Press, 1978.

Edelman, G. M. Neural Darwinism: The Theory of NeuronalGroup Selection. New York: Basic Books, 1987.

Edelman, G. M. The Remembered Present: A Biological Theoryof Consciousness. New York: Basic Books, 1989.

Edelman, G. M. Bright Air, Brilliant Fire: On the Matter ofthe Mind. New York: Basic Books, 1992.

b i b l i o g r a p h i c n o t e182

Edelman, G. M. “Neural Darwinism: The Theory of Neu-ronal Group Selection.” Neuron 10 (1993): 115–125.

Edelman, G. M., and Tononi, G. A Universe of Consciousness:How Matter Becomes Imagination. New York: Basic Books,2000.

For an account of the concept of degeneracy, see:

Edelman, G. M., and Gally, J. A. “Degeneracy and Complex-ity in Biological Systems.” Proceedings of the National Acad-emy of Sciences USA 98 (2001): 13763–13768

For two recent papers considering various aspects of ascientifically based approach, see:

Crick, F., and Koch, C. “A Framework for Consciousness.”Nature Neuroscience 6 (2003): 119–126.

Edelman, G. M. “Naturalizing Consciousness: A TheoreticalFramework.” Proceedings of the National Academy of Sci-ences USA 100 (2003): 5520–5524.

This last paper succinctly puts forth the views elaborated inthis book. For related references to neural correlates of con-sciousness, the following volume is an excellent source:

Metzinger, T., editor. Neural Correlates of Consciousness: Em-pirical and Conceptual Questions. Cambridge: MIT Press,2000.

For works bearing on the experiments described inChapter 9, see:


Leopold, D. A., and Logothetis, N. “Activity Changes inEarly Visual Cortex Reflect Monkey’s Percepts DuringBinocular Rivalry.” Nature 379 (1996): 549–553.

Tononi, G., Srinivasan, R., Russell, D. P., and Edelman,G. M. “Investigating Neural Correlates of Conscious Per-ception by Frequency-Tagged Neuromagnetic Responses.”Proceedings of the National Academy of Sciences USA 95(1998): 3198–3203.

Srinivasan, R., Russell, D. P., Edelman, G. M., and Tononi,G. “Increased Synchronization of Neuromagnetic Re-sponses During Conscious Perception.” Journal of Neuro-science 19 (1999): 5435–5448.

Finally, a set of accounts with which I disagree shouldbe included for balance and fairness. The modern authorslisted below find themselves in the distinguished company ofRene Descartes.

Descartes, R. The Philosophical Works of Descartes, 2 vols.,edited by E. Haldane and G. Ross. Cambridge, England:Cambridge University Press, 1975.

Popper, K., and Eccles, J. F. The Self and Its Brain. NewYork: Springer, 1977.

Penrose, R. Shadows of the Mind: A Search for the MissingScience of Consciousness. New York: Oxford UniversityPress, 1994.

McGinn, C. The Problem of Consciousness: Essays Toward aResolution. Oxford: Blackwell, 1996.

Chalmers, D. The Conscious Mind: In Search of a Fundamen-tal Theory. New York: Oxford University Press, 1996.


If an insatiable reader wishes an even longer list of refer-ences, I refer him or her to David Chalmers’s annotated com-pendium on the World Wide Web:

http://www.u.arizona.edu/�chalmers/biblio.html

The exploding list of references speaks to the conclusion thatthe understanding of consciousness has a promising scientificfuture.


187

Index

acetylcholine, and cholinergicnuclei, 25, 152

action potential, 18, 149adaptive immune system, 131–

132, 149algorithms, and computation,

38, 111, 153alien hand syndrome, 136amino-acid sequences, and

genetic code, 43, 160amplification, and selectionism,

42amygdala, 26animals: and higher-order con-

sciousness, 134, 161–162;and phenomenal transform,81; primary consciousness of,11, 56, 57–58, 59, 77, 97,134

anosognosia, 38, 136, 143,149, 153

anterior commissure, and dis-connection syndromes, 143

aphasia, 100, 149, 180ascending systems. See value sys-

temsassociation areas, 89, 150associativity, 150; and con-

scious experience, 135–136;and degenerate interactions,114, 126; and language ac-quisition, 103, 136

atomism, logical, 107attention, 150; and basal gan-

glia circuitry, 88, 90, 94, 95,127, 143; and conciousstates, 7, 94–95, 143; motorcomponents, 95; and non-conscious execution, 88, 93–94; and TNGS, 127–128.See also focal attention

auditory discrimination, 102

i n d e x188

automaticity, 88, 127–128,144–145, 150

automatic motor processes, 93,144

autonomic nervous system, 57,73, 128–129, 133, 150

axons, and synaptic connec-tions, 17, 18, 150

basal ganglia, 16, 21, 53, 88–89, 150, 170; and atten-tional states, 143; and auto-maticity, 127–128; motorcircuitry of, 88–96; motor se-quencing and control, 24,103; polysynaptic loop struc-ture of, 24, 26–28, 96, 163;and surround effects, 126;thalamocortical system and,69, 70–71

binding problem, 36, 44, 150–151

binocular rivalry, 108, 110,151

bipedal posture, 102, 103, 151blindsight, 143, 151blind spot, and brain dynamics,

37, 136brachiation, and gestural com-

munication, 102brain: anatomy and dynamics

of, 14–23, 26, 27, 151; andascending value systems, 25–

26, 28; capacity to general-ize, 38; computer model of,28–30, 33, 35–36, 39, 45–46, 114; and integration,136; motor functions, 23–24; as origin of conscious-ness, 4–5, 6–7, 21; sensoryand cognitive functioning of,14–23

brain scans, 151. See also mag-netoencephalography (MEG)

brainstem, 16, 151; and catego-rization, 72–73; and modula-tion of emotion, 133

Brentano, Franz, 125, 164Broca’s area, of cerebral cortex,

99–100, 101, 151–152Bruner, Jerome, 100

C, as conscious process, 78–86,130, 152; and attention,127–128; and integration,121; relationship with C′(neural activity), 115–116,117, 122–123, 145–146; so-cially defined self and, 137–138

C′, as neural activity, 78–86,130, 152; and attention,127–128; centrality of, tocomplex subjective states,130; and looped reentrant in-teractions, 122–123; relation-

i n d e x189

ship with C (conscious pro-cess), 115–116, 117, 122–123, 145–146; and sociallydefined self, 137–138

categorization. See perceptualcategorization

caudate nucleus, 89causal efficacy, 75, 86, 152causality, and conscious pro-

cess, 76–78, 81–82, 152cell body, 17, 152cell migration, 28–29, 152cerebellum, 16, 21, 96, 152;

and automaticity, 128; cir-cuitry of, 89, 95; motor con-trol and sequencing, 23, 88,89; and temporal succession,53, 88

cerebral cortex, 15–17, 83,152, 154; and basal ganglia,89–92; development andevolution of, 102; hemi-spheres, 20, 154; and mem-ory, 51, 88; motor areas of,91, 92, 93; and thalamocor-tical system, 26–28, 54, 69,70, 71, 88–89. See also so-matosensory cortex; Wer-nicke’s area, of cerebralcortex

channel, 17–19, 152cholinergic nuclei, 25, 93, 133,

152

closure (filling-in), and braindynamics, 120, 136, 153

coarticulated sounds, 102, 153codons (triplets), and genetic

code, 43, 160cognitive defects, and basal gan-

glia, 93cognitive science, and “represen-

tation,” 103–105, 111coherency, and reentry, 45,

153color experiences, and qualia,

64–65complex systems, and integra-

tion, 65–68, 153computations, 105, 153computer models, 41, 153; of

brain and mind, 28–30, 33,35–36, 39, 45, 107, 114;and conscious process, 85–86

concepts, 153–154; formationof, 50, 53; and mapping ofuniversals, 104–105

consciousness: of consciousness,101, 116; and discrimina-tion, 10, 72–73, 141–142;experience of, 4–13; and in-tegration, 7–8, 136; neuralcorrelates of, 107–112, 118;perceptual categorizationand, 72–73; and self-awareness, 73, 116; structure

i n d e x190

consciousness (continued)and dynamics of, andTNGS, 141–145. See also C,as conscious process; C′, asneural activity; higher-orderconsciousness; primary con-sciousness

conscious state, and TNGS:general properties of, 119–125; informational func-tions, 119, 120, 125–128;and irreducibility, 124–125;subjective properties, 119,120, 128–130; summaryand testability of, 110, 113–118

corpus callosum, 20, 143, 154cortex. See cerebral cortexcorticocortical tracts, 20corticostriatal projections, 92,

154corticothalamic projections, 20

Darwin, Charles, 1–3; The De-scent of Man, 2; and naturalselection, 32–33, 47, 157

Darwinism, neural, 33, 147.See also neuronal group selec-tion, theory of (TNGS)

degeneracy, 43–46, 154; anddynamic core, 70, 72, 121;and genetic code, 160; ofneural networks, 44–45, 52,

111, 114–115, 147; re-entrant circuitry of, 46, 106,121, 135; selective processes,114; and synaptic response,51, 52

dendrites, of neurons, 17, 18,154

Descartes, Rene, 2, 155–156,174

The Descent of Man (Darwin),2

developmental selection, andTNGS, 39–40, 155

differentiation, of conscious ex-perience, 31, 106, 128, 155

digital computation, and brainaction, 29–30. See also com-puter models

disconnection syndromes, 143discrimination, 155; adaptive

value of, 135; auditory, 102;and qualia, 70, 72, 146; andreentrant interactions, 141–142; and subjectivity, 131

dissociative syndromes, 143distributed systems, 36, 41,

155diversity, and selectional sys-

tems, 41–42dopamine, as neurotransmitter,

24, 25, 92, 155dopaminergic projection, from

substantia nigra, 92

i n d e x191

dopaminergic system, 133dualism, 137, 145, 155–156,

174dynamic core, 156; adaptive be-

havior and, 132–133; andcategorization, 74; and con-sciousness, 84–85, 96, 115,127, 143; cortex interac-tions, 74, 94; as functionalcluster, 69, 74, 96, 156,159; modulation of, 127–128, 142–143; and phenom-enal transform, 78–81; re-entrant interactions, 69–72,77–78, 79, 85, 96, 143

econiche, 136, 156, 176effective procedures, and com-

puter models, 35embodiment, 5, 156emotion, 156; and amygdala,

26; and self-discrimination,129; and value systems, 130,133

endogenous opioids, 26, 156enkephalin, 26, 156entailment, and relationship of

C and C′, 80, 81, 85–86,146, 156

entropy, informational, 66–67,157

epigenetic processes, 29, 157epiphenomenonalism, and con-

scious process, 82, 85, 137,145, 157

episodic memory, 22–23, 51,88, 157, 162

evolution, 157; and brain devel-opment, 32–33; of consciousprocess, 54–57, 85, 138–139; of primary conscious-ness, 57, 132; as selectionalsystem, 32–33, 41; value-system constraint, 42, 138–139

evolution, theory of, 157; andbiological basis of conscious-ness, 1–3; and TNGS, asneural Darwinism, 41–43,157

excitatory inputs, 25, 89–90experiential selection, and

TNGS, 39, 114, 157–158.See also selectionism

explanatory gap, 11–12, 158eye movements, and construc-

tive “filling-in,” 126–127

feedback, 41, 158filling-in, as neural dynamic,

116, 120, 122, 124, 127first-person perspective, 63,

74–75, 158fMRI (functional magnetic reso-

nance imaging), 151, 158,166

i n d e x192

focal attention, 158; and con-scious state, 7, 61, 120; anddynamic core, 94–95, 127,142–143

foveal discrimination, and eyemovement, 126–127

frequency tag, 159Freudian unconscious, 95,

159fringe effects, and conscious

state, 7, 126–127, 159functional cluster, 69, 96, 156,

159functional connectivity, 67, 68,

159functional magnetic resonance

imaging (fMRI), 151, 158,166

functional segregation, 68,159

GABA (gamma-aminobutyricacid), as neurotransmitter,19, 24, 93, 159–160

Gall, Franz Joseph, 30, 171genetic code, and degeneracy,

43, 160gestalt phenomenon, 160; as

feature of conscious state,116, 120; and synchrony ofreentry, 124

gestural communication, 102,160

glia (support cells), 29, 160global brain theory, 33. See also

neuronal group selection, the-ory of (TNGS)

global mapping, 160, 176;inhibitory output of, 95,127; and perceptual cate-gorization, 49–50, 53, 56,144; and TNGS, 114–115

globus pallidus, 89, 90, 92,160–161, 170

glutamate, as neurotransmitter,19, 92, 161, 169

glutamatergic inputs, 25, 89–90

gray matter, of brain, 16, 152

Hayek, Friedrich von, 22Hebb, Donald, 22, 161Hebb synapse, 22, 161hemineglect, as neuropsycholog-

ical syndrome, 136, 143,161

Heraclitean illusion, 103, 134,161

high-dimensional space, 135,161. See also qualia space

higher-order consciousness,8–9, 58–59, 161–162; evo-lution and development of,73, 101; experiments andmagnetic field testing, 107–111; linguistic ability and,

i n d e x193

9, 58, 100, 101, 102–103,112; and representation,97–112; self-concepts of,77, 129–130, 134, 137–138; TNGS theory, 116–130

hippocampus, 16, 53, 157,162; and episodic memory,22, 51, 88, 99; synapticmechanisms of, 21–23

histaminergic system, 25holistic view, of brain function,

30–31, 168homeostatic system, 57, 133,

162hominines, and higher-order

consciousness, 58, 98, 100,102

homologous structures, 59,132, 162

homunculus, 46–47; and expe-rience of conscious process,85, 86; as observer, 74–75

Hox genes, 29, 162Huntington’s disease, 93, 162Huxley, T. H., 82–83, 84hypnotic trance, 143hypothalamic systems, and emo-

tional response, 133hypothalamus, 25, 150, 162–

163hysteria, as dissociative syn-

drome, 143

ideal (perfect) gas, 66, 67, 163identity, 131–139, 163illusion, 163; phenomenology

of, 36–37; and reentrant con-nections, 122, 136

imagery, and brain dynamics,105, 126, 144

immune system, 41, 42, 44,131, 149

individual variation, and evolu-tion, 132

individuation, and self-discrimination, 129. See alsoself-concept

informational properties, 120,125, 163

inhibitory loops, structure of,and basal ganglia, 24, 26–28, 96, 163

instructionism, 41, 163, 175.See also logic; selectionism

integration, 164; brain activityand, 31, 106, 136; and com-plex systems, 65–68; of con-scious state, 121, 122; andhigher-order consciousness,135

intentionality, and conscious-ness, 105–106, 120, 125,164

intralaminar nuclei, of thala-mus, 21, 31, 89, 164

irreducibility, 124–125, 164

i n d e x194

James, William, 4–7, 55, 82,83–84, 134, 172, 174;“Does Consciousness Exist?”6; The Principles of Psychol-ogy, 4, 82–83, 177

Jamesian properties, 7, 31, 164

Kanizsa triangle, 36–37, 163kinesthetic response, 57, 164;

and motor control, 133; andself-perception, 72–73, 129

language, 164–165; andhigher-order consciousness,58, 98, 112; origins and de-velopment of, 99–103, 135

Lashley, Karl, 34lateral geniculate nucleus, of

thalamus, 20, 165, 170lexicon, 98–99, 103, 118, 165linguistic ability, 165; develop-

ment of, 100–103, 135–136; and higher-order con-sciousness, 9, 118

locus coeruleus, and ascending-value system, 25, 26–28,133, 165

logic: as organizing principle,41; and selectionism, 147–148

logical atomism, 165long-term memory, 23, 88, 99,

165

magnetic resonance imaging(fMRI), 151, 158, 166

magnetoencephalography(MEG), 107–109, 110, 118,151, 159, 164, 166

maps, 40, 41, 44–45, 49–50,53, 100, 104, 114–115,122, 136, 166

meaning, in neurobiology, 105,166. See also semantic capa-bilities

memory, 50–53, 166, 176;hippocampus and, 21–22,51, 99; as nonrepresenta-tional, 52–53, 104–105,115; and perceptual categori-zation, 114–115, 129–130;reentrant interactions, 124;sensory and motor modula-tion, 115, 124, 143, 144.See also episodic memory;long-term memory; proce-dural memory; short-termmemory; value-categorymemory

mental images, 9, 104, 105,167

mental representations, 103–107, 109, 167

metaphor, 167; and higher-order consciousness, 103,135–136

metastable, 74, 115, 126, 167

i n d e x195

milliseconds, and neural re-sponse, 144, 167

mime, as gestural communica-tion, 102, 167

mind, neural basis of, 1–3,140–148, 167

modularity, and brain function,30–31, 167–168

modulation, 168; of dynamiccore, 127, 142–143

motor control: and basal gan-glia, 24, 90–93, 127; andconcepts, 105; and consciousattention, 61–62, 90, 95,144–145; global mapping,114–115, 144; language ac-quisition, 102–103

motor regions, of the brain,23–24, 88, 91, 92, 168

mutual information, 66, 163,168

natural selection: and degener-acy, 44; and global brain the-ory (TNGS), 2–3, 32–33;and process of consciousness,48; as selectional system, 41,42

natural selection, theory of, 1–2, 47, 168

nervous systems, and naturalselection, 4, 41, 42

neural correlate of conscious-

ness, 13, 59, 107, 118, 137,145, 168

neural Darwinism, and TNGS,32–47, 147, 168

neural dynamics, and reentry,116

neuroanatomy, and complexsystems, 67

neuromodulators, 25–26, 168–169

neuronal groups, 169; degen-eracy of, and memory, 52;and reentry dynamics, 123–124

neuronal group selection, the-ory of (TNGS): as globalbrain theory, 33–39, 45–47,179; and memory, 50–53,114–115; reentrant interac-tions, and degeneracy, 39–41, 43–45, 46, 176; tenetsof, and selectionism, 39–45,53–54, 72, 84–85, 114,155–157, 179. See also con-scious state, and TNGS

neurons, 14, 15, 16–19, 169neuropeptides, 25–26, 168–

169neurophysiology, 109, 111,

169neuropsychological syndromes,

and illusory phenomena,37–38, 143

i n d e x196

Neurosciences Institute, 118neurotransmitters, 17–19, 24,

169noise, and brain function, 35,

114, 169nonconscious brain activity,

88, 93–94, 95–96, 143,169

nonrepresentational memory,52–53

nonself, 129, 170noradrenaline, 25, 165nuclei, as clusters of neurons,

20, 170

obsessive-compulsive disorders,93

optic nerve, 20, 37, 170

parietal cortex, 17, 91, 94, 161Parkinson’s disease, 24, 92–93,

170pars reticulata, of substantia ni-

gra, 90, 170pattern recognition, 38–39,

147. See also selectionismPax genes, 29, 162perception: and context, 36–

38; motor elements of, 144perceptual categorization, 49–

50, 117, 160, 170; andconsciousness, 55–57,72–73, 114–115, 125,

143, 144; evolution of, 54,55, 57; and memory sys-tems, 50–51, 53, 55–57,121, 129–130; and re-entrant interactions, 49–50,54, 56, 57, 121; and TNGS,50–51, 114–115, 117,125

perfect crystal, 66, 68, 170phenomenal experience, or con-

sciousness, 61–63, 170phenomenal transform, 77–78,

117, 170–171; and neuralcore processes, 78–80, 86;and reentrant interactions,123; and self-reference, 128–129

phenomenology of illusions,36–38

phenotype, properties of, 146phonetics, 171phrenology, and brain faculties,

30, 171Piaget, Jean, 171Piagetian notion of self, 129,

171polysynaptic loop structure, of

basal ganglia, 24, 26–28, 96,163

population thinking, 33–35,171

postsynaptic neurons, 17, 18,20, 22, 171

i n d e x197

prefrontal cortex, and cognitivedefects, 93

premotor regions, 50, 89, 92,171

presynaptic neurons, 17, 18,22, 172

primary consciousness, 8–9,48, 56–59, 77, 97, 172,180; in animals, 63–64, 77,97; evolutionary path of,132, 133; links with higher-order consciousness, 59, 99,116–118; mental images in,105, 167; and “rememberedpresent,” 118, 134, 144; andself-concept, 77, 97, 132,133–134; and TNGS, 116–118

primary motor cortex, 23, 154primates, and higher-order con-

sciousness, 98. See also homi-nines

The Principles of Psychology( James), 4, 82–83, 177

privacy, 75, 172procedural memory, 24, 51,

150, 172progenitor cells, 28–29, 172propositional attitudes, 119,

172proprioception, 57, 172; and

motor control, 133; and self-systems, 73, 128–129

prosopagnosia, 143, 172–173

protosyntax, 103, 173putamen, of basal ganglia, 89,

173

qualia: and causality, 152; andhigher-order consciousness,135; and high-order discrimi-nation, 64–65, 85, 115,117, 125, 142, 146; and phe-nomenal transform, 62–63,77–78, 86, 117; and pri-mary consciousness, 10–11,73–74, 115; and representa-tion, 106; as subjective con-scious state, 3, 10–13, 120,173

qualia space: and discrimina-tory capabilities, 73, 115,128, 129; and integration,72; and phenomenal trans-form, 86

raphe nucleus, 25, 133, 173rapid eye movement (REM)

sleep, and conscious states,9–10, 122, 144, 176

recategorical memory, 52, 173receptors, 18–20, 173, 176reciprocal fibers, 20, 39, 173reentrant mapping, and TNGS,

39–41, 43, 136

i n d e x198

reentry, 110, 114, 116, 121,141, 143, 174; concatenatedloops, 122–123, 153; degen-eracy of, 45–46, 106, 135;and dynamic core, 69–72,79, 95, 121, 143; “looped”or cyclic nature of, 122–123; measurement of, andMEG, 107–109, 110, 118;and primary consciousness,56; and spatiotemporal coor-dination, 41, 45, 100, 114;thalamocortical system and,54–55, 68–72

reflex, 174“remembered present,” 8, 55,

103, 118, 134, 144, 174repertoire, 33, 39, 40, 42, 111,

114, 174representation, 174; and con-

sciousness, 104–107; andMEG experiment, 107–109,110; and reentrant patterns,109, 111–112

repression, and Freudian uncon-scious, 95

res cogitans, 2, 156, 174res extensa, 156, 175reticular nucleus, of thalamus,

21, 54–55, 68, 94, 127,175

retina, and eye movements,126, 175

Russell, Bertrand, 165

scene, 7–8, 11, 48, 55–58, 61,115, 116, 175

schizophrenia, 10, 155, 175selectionism, 157–158, 168,

175; and global brain the-ory, 32–33; and relation-ship with logic, 147–148;and TNGS, 39, 41–43,114

self-concept, 175; characteris-tics and development of,132–134; and consciousness,74–75, 77, 134, 175; anddynamic core, 77, 132; andhigher-order consciousness,101, 116; Piagetian notionof, 129; as process, 137–138; and self-discrimination,128–130; and semantic capa-bilities, 77, 116, 118

semantic capabilities: andhigher-order consciousness,9, 58, 77, 98, 99, 116, 118;origins of, 99–102

semantic memory, 176semantics, 166, 176sensorimotor loops, 176. See

also global mappingsensory modalities, and cerebral

cortex, 124sensory receptors, 176

i n d e x199

sequencing, and basal ganglia,24

serotonin, 25short-term memory, 51, 53,

88, 99, 176sleep, and consciousness, 9–10,

176. See also rapid eye move-ment (REM) sleep

social development, and indi-viduation, 116, 129–130,137

social values, 138somatoparaphrenia, 136, 176somatosensory components, of

“self ” system, 73somatosensory cortex, 17, 35–

36, 91spatiotemporal coordination,

41, 114spatiotemporal correlation, 176“specious present,” 55, 177speech community, 137, 175,

177Sperry, Roger, 84split-brain syndrome, 20SQUIDS (superconducting

quantum interference de-vices), 166

stochastic processes, 29, 177striatum, of basal ganglia, 89,

90–92, 177subcortical structures, 87–88,

100, 177

subcortical value systems, 68subjectivity, 177; and conscious

states, 63, 120; and self-reference, 128–130, 131,134

substantia nigra, 89, 90, 92,170, 177

subthalamic nucleus, 89, 90,177

supervenience, 81–82, 178supralaryngeal space, 102,

178surround effects, and conscious

state, 124, 126–127synapses, 15, 16–19, 178;

and developmental selec-tion, 35, 39, 54; strengthand efficacy of, 22–23,178

synaptic response, and memory,22, 50–52, 54

synaptic vesicles, 18, 178synchrony, and reentrant cir-

cuits, 41, 45, 114, 124,178

syntactic capabilities, 98, 99–103

syntax, 178

thalamic nuclei, 21, 23, 31, 54,177

thalamocortical maps, and tem-poral succession, 53

i n d e x200

thalamocortical projections, 20thalamocortical system, 16,

26–28; and basal ganglia,88–96; evolutionary changesin, 54; and primary con-sciousness, 115; reentrantdynamics of, 68–72, 95, 96

thalamus, 177, 178–179;anatomy and dynamics of,19–21; and attentionalstates, 143; and basal gan-glia, 88, 89, 90, 91, 170;evolutionary changes in, 54–55; intralaminar nuclei, 21,31, 54, 179; and reentrantconnections, 54; and somato-sensory cortex, 23, 35–36,55. See also thalamocorticalsystem

third-person perspective, 74–75, 179

TNGS. See neuronal group se-lection, theory of (TNGS)

triplets, and amino-acid se-quences, 43, 160

Turing, Alan, 179Turing machine, 33, 84, 153,

179

unconscious, Freudian, 95unconscious state, 9–10, 179.

See also nonconscious brainactivity

unitary scene, 7–8, 10, 61,64–65, 68, 115, 116, 123,128, 135, 179

value-category memory, 53–54, 117, 180; and percep-tual categorization, 55–57;and primary consciousness,56, 57, 58, 115, 180; andreentrant interactions, 56,57, 115, 121, 180; self-discrimination and, 132–133

value systems, 25–28, 113,180; and amplification,42; and emotional response,133; evolutionary selectionand, 42, 113–114, 138–139; and perceptual cate-gorization, 72–73, 125;and population think-ing, 34–35; and self-discrimination, 128–129,130. See also value-categorymemory

variability, 113–114, 180veridical reality, 135, 136,

180vision, and “fringe” effects,

124, 126–127visual cortical areas (V1, V2,

V3, V4, V5), 20, 30–31, 68,124, 165, 179

i n d e x201

visual perception, and imagery,126

Wallace, Alfred, 1–2Wernicke’s area, of cerebral cor-

tex, 99–100, 101, 180

Wittgenstein, Ludwig, 106,165

zombie hypothesis, 180; andC–C′ entailment, 80, 145–146

Documents

[Dr. Gerald Edelman Ph.D.] Wider Than the Sky the(BookFi.org)