I am indebted for advice, much of it heeded, from M. Adelson, G. M. Bogen, R. W Doty, D. Galin, C. Hamilton, J. Johnstone, F. Michel, M.O’Dell, C. Trevarthen, A. van Harreveld. A. Vaughan, R. Zachariasen, and E. Zaidel. Thanks are due for word processing to Sally Johnstone and for library assistance from S. Zeind, Chief Librarian of the Huntington Memorial Hospital in Pasadena.

1. INTRODUCTION

This paper is a followup to my articles called "The Other Side of the Brain" which were published in the Bulletin of the LA Neurol. Soc. in 1969 (24). Those articles presented ideas some of which may now seem fairly orthodox. But in those days they were not, and I think that for Dr. Richard Walter , then editor of the Bulletin to publish them was a tribute to his wry sense of humor, as well as his historical perspective.

I shall now present evidence accumulated since 1969 that the onebrain view is inadequate to describe not only the "split-brain" but also the anatomically intact encephalon. I shall consider in some detail the possibility of retaining both the single brain and double brain views, as an example of complementarity, a philosophical concept adopted by some quantum physicists. Hence, a few philosophical ideas implicit in the remainder of the paper should be mentioned.

Does an individual having two cerebral hemispheres have, in some important sense, two brains rather than only one? This "important sense" deserves to be made more explicit at the outset. It includes, I believe, the following: that whatever may be for neurologists the meaning of the psychological term "mind," the number of brains in an individual is the same as the number of minds. If this equivalence of numerosity were not agreed upon, we might well spend the rest of our time arguing that question as well as others, such as: What meanings are possible for the word "mind"? Without demeaning such perennial questions, I will take it for granted, at least for this paper, that what we can learn about mind also tells us about brain and vice versa. In other words, the mentalistic and physiologic are two different descriptions of the same underlying reality.

When discussing the duality versus singularity of brain, or in shifting back and forth between mentalistic and physiologic vocabularies, we are not concerned with a quite different issue, one which is often couched in misleadingly similar terms. That is, we are not here concerned with the metaphysical issue of dualism vs monism, whether the monism be materialistic or idealistic. Whether one professes (as have so many Eastern philosophers and some contemporary quantum physicists) (68) that mind begets brain, or whether one supposes [along with Sir Charles Sherrington, (9)] that mind coexists with and interacts with brain, or whether one supposes (as I believe) that mind is generated by brain, need not concern us here. Our metaphysical views neither entail nor are they entailed by our views on the question at hand. This is because one could easily be on either side of the dualism /monism issue irrespective of one's position on the question of duality or singularity of mind.

How is it that this question (duality versus singularity of brain) has come so forcefully to our attention in the past few years?

The view that the brain should be considered not as a single organ but rather as a pair of organs, in much the same sense that each of us has a pair of kidneys or a pair of lungs, is at least as old as the writings of Hippocrates. In his words, ". . . the human brain, as in the case of all other animals, is double" (10). This view was urged with utmost conviction by Arthur Ladbroke Wigan, an English physician who published his book, The Duality of the Mind, in 1844. Among the many other similar sentiments in this book, he claimed that, "the mind is essentially dual like the organs by which it is exercised." He believed himself able to prove, "that a separate and distinct process of thinking or ratiocination may be carried on in each cerebrum simultaneously" (11). Today we call this the concept of cerebral duality or hemispheric independence; it is important to understand that this proposition is not dependent upon the existence of hemispheric specialization; it is true of cats, monkeys, etc., as well as humans.

The view of the brain as double was also advanced by others, notably including Brown-Sequard, David Ferrier, and Sir Victor Horsley (3). John Hughlings Jackson, in his essay entitied "On the Nature of the Duality of the Brain," wrote as follows:

"That the nervous system is double physically is evident enough. This is a very striking fact, but one so well known that we are in danger of ceasing to think of its significance or of ceasing to wonder at it. The nervous system, I repeat. is physically double. I wish to show that it is double in function also, and further, in what way it is double in function" (12).

Although such views were quite popular in the late 19th century (13) the view that the brain is double was in abeyance for nearly half a century. It has only recently been revived in the light of the splitbrain results, both animal and human (1416). That this view remained in eclipse for so long was probably of multiple causation, but I would suggest that perhaps the strongest of these causes has its origin not in objective science but in subjective intuition-the inner conviction of wholeness, or unity, of singleness felt by each of us.(Itis sometimes acclaimed as "the unity of consciousness"). According to Rene Descartes:

". . . there is a great difference between the mind and the body, in that the body is, by its nature, always divisible, and the mind wholly indivisible. For, in fact, when I contemplate it-that is, when I contemplate my own self-and consider myself as a thing that thinks, I cannot discover in myself any parts, but I clearly know that I am a thing absolutely one and complete"(17).

The same conviction was elegantly expressed by Sir Charles Sherrington when he wrote:

"This self is a unity . . . it regards itself as one, others treat it as one. It is addressed as one, by a name to which it answers. The Law and the State schedule it as one. It and they identify it with a body which is considered by it and them to belong to it integrally. In short, unchallenged and unargued conviction assumes it to be one. The logic of grammar endorses this by a pronoun in the singu.lar. All its diversity is merged in oneness." (9).

However correct or erroneous it may be, it is this conviction which offers the greatest resistance to the twobrain view, even now that the splitbrain results are so widely appreciated (18, 19). One can still see in current print, for example, such sentiments as "the brain is a unity" or "all parts of the brain participate in all of its actions" or [and this may be the most familiar form] "the brain works as a whole." This is not a description of an observable fact; it is the sloganof a particular philosophic conviction. [There is a sense in which it might be considered consistent with the evidence. That is, we could say that the cerebrum works as a whole in the sense that every brain cell has a certain probability of contributing significantly to any specified function or bit of behaviour . In this way it could be said that a neuron in the right temporal pole (of a right hander) has a certain probability of being necessary for speech (but a probability vanishing small!).

The subjective evidence, that is to say introspection, recurrently reminds each of us that "mind is single.'' But the objective, experimental evidence increasingly tells us that the intact brain is double. This objective evidence is the main content of this paper.

2. HISTORICAL BACKGROUND

This paper offers a thesis which can be outlined in three sentences. The first is, "One cerebral hemisphere is enough for a mind": the best evidence for that comes from hemispherectomy. The second statement is, "with two hemispheres it is possible to have two minds." The best evidence, or at least the most striking (there is quite a lot of other evidence) is from the human splitbrain, where the two minds appear to be so different.

The third assertion is that this duality of mind can be present in the intact cerebrum (5, 20). I will present some of the evidence and some of the arguments for these three statements-mainly for the third statement because the first two are by now fairly well accepted.

It was about 1820 when Arthur Ladbroke Wigan lost an acquaintance from a relatively sudden death (possibly a heart attack). At the postmortem, when the head was opened, he and the others present were astonished to see that one hemisphere was completely missing. He was not only astonished, but he recognized that this was meaningful, so he looked around for other cases and he found a few. In his book The Duality of the Mind published over 20 years later, he took it as unarguable that one hemisphere is enough for a mind (11). There have been naturally occurring resorptions of a single hemisphere observed since, but the best evidence, because it is so readily replicable is from hemispherectomy. The evidence comes first of all in animals. I won't cite all of the experimental observations, having reviewed most of them previously (21); but I would like to point out what Wenzel, Tschirgi, and Taylor said about their experience with hemispherectomized kittens. In spite of removal of nearly onehalf ofthe brain, "the behavioral loss is extremely small, or even undetectable" (22).

Perhaps even more impressive than their informal behavioral descriptions are the formal, quantitative investigations by physiological psychologists. There have been many such studies. One that is especially notable, because of the many years it continued, is the research of Kruper , Koskoff and Patton (23). Patton observed in 1961 that the hemicerebrectomized monkey can achieve a level of performance on a complex learning taskwhich equals that of the normal animal, although it may take a little longer. Ten years later, these authors wrote,

". . . showing clear learning defcits when comparing the performance of these animals to the performance of normal monkeys continues to be a compelling challenge" (23).

The quality of mind which the residual hemisphere affords depends on the integrity of the residual hemisphere that is, what and and how much damage it has had. If "mind" includes (at least) the abilities to perceive, discriminate, compare with previously stored information, consider alternative actions, choose among them, and then to act, it is doubtful that any smaller subdivision of a mammalian brain can support a mind. That is, although these minimal criteria for "mind" do not distinguish what a hemisphere can do from what the best electronic artifacts can do, it does eliminate any favorite gyrus, cerebral lobe, or cord segment from consideration as the seat of a "mind." By contrast, one hemisphere properly connected can support a mind, by these or any other familiar criteria. Not only neuroscientists but many laymen are by now familiar with these facts. For example, C. E. Marks (a philosopher at the University of Washington in Seattle) wrote that no one today doubts Wigan's original point that one hemisphere is enough for a mind "comfortably characterizable as human" (24).

If one hemisphere can be enough for a mind, can two hemispheres be enough for two minds? When they are in separate heads, the answer is clearly "yes." Two people, each with a hemispherectomy, have two minds. But what if the two hemispheres are in the same head? Most readers already have some familiarity with the splitbrain phenomena. It requires emphasis that the concept of duality of mind is not based only on the humans who have had a splitbrain operation. The human cases may be more dramatic, but the duality of mind when the brain is split was established before there were any human cases, in hundreds of experiments by dozens of investigators, in the cat and monkey (25). Of course, the human cases attracted more attention, Of the human cases ,this is what Roger Sperry said in 1974:

"Although some authorities have been reluctant to credit the disconnected minor hemisphere even with being conscious, it is our own interpretationóbased on a large number and variety of nonverbal testsóthat the other hemisphere is indeed a conscious system in its own rightóperceiving, thinking, remembering, reasoning, willing, and emotingóall at a characteristically human level, and that both the left and right hemisphere may be conscious simultaneously in different. even mutually conflicting. mental experiences that run along in parallel "(15).

We come now to the third point. What about two hemispheres inside one head when the two hemispheres are not anatomically disconnected?

To what extent does the "duality of mind,'' easily demonstrable in the splitbrain, also obtain in the normal individual? This is a principal question affecting most of the inferences, regarding human nature, from the splitbrain data. Indeed, hemispheric specialization would be of little more interest than the many other differences in cortical representation of function, if it were not for its concurrence with a significant degree of hemispheric independence.

To have any duality of mind, hemispherically based, there must be some circumstances in which memory traces are unihemispheric and are not communicated to the other hemisphere, even with commissures intact. This was considered by Doty et al. in their article on "The Unilateral Engram," in which they went so far as to suggest that the corpus callosum may actually act to restrict memory storage to one or the other hemisphere (26). I shall not take so strong a position here, being content only to claim that the neocommissures cannot affect a complete unification of information processing in the two hemispheres.

There are two aspects to my claim: first, that without the neocommissures there are nevertheless many effectively unifying mechanisms, and second, that even with these plus the neocommissures. unification is incomplete, resulting in mental duality, hemispherically based. We consider first some of the extracallosal sources of unification.

3. SOME EXTRACALLOSAL UNIFYING MECHANISMS

That there must be some synchronization (with or without the commissures) is clear enough. Unlike two hemispheres in two different heads, the disconnected hemispheres have been and continue to be in the same place at the same time, have the same circulating environment (blood and CSF), and they communicate a great deal subcallosally. This subcallosal sharing includes their possession in common of the "second" (midbrain dependent) visual system (29, 30). Brain stem connections make for unification in other ways. For example, the duality of cognition demonstrable in the human splitbrain seems not to be accompanied by so great a disparity of affect (31, 32), perhaps because the two limbic systems are so tightly coupled at the hypothalamic level.

An important source of synchronization is the brain stem activating system, particularly since generalized arousal via this system can be produced by descending corticofugal influences from one hemisphere (33). Both human hemispheres, whether in the intact or the split condition, are probably asleep and awake simultaneously, just as they are in the splitbrain cat (34, 35) unless there is a sagittal section of the brain stem (36, 37). There is no evidence that primate hemispheres (in the same head) can alternate sleep and waking, as do cetacean hemispheres (384]).

lf it were the case that the two hemispheres in normal continuity were some times independently awake and asleep, that is, if wakefulness varied independently, then cognition could also vary independently. In the dolphin (whose hemispheres are well connected by a large corpus callosum) this appears to be the case. If the dolphin hemispheres can sometimes be alternately awake (hence independently cognizing when anatomically intact, it suggests widely varying degrees of the synchronizing capacity not only of the brain stem but also of the corpus callosum.

Before reviewing some of the evidence for partial hemispheric independence in mammals other than cetaceans, we should first consider some other reasons why hemispheric independence is incomplete in the splitbrain. The unifying function of the corpus callosum has been unduly emphasized, not only when it is intact (the gist of this paper) but it has also been overemphasized also when no longer present.. As noted above, there is a lot ofsubcortical commumcation via the brain stem whether or not the neocommissures are cut. And the cerebellar commissures, less often considered in this context may well contribute, especially in the absence of the (probably) overriding callosal transmission. Moreover, if one hemisphere initiates motion, the other hemisphere receives considerable proprioceptive feedback. Since the hemispheres share a common blood supply, if one hemisphere causes the secretion of some hormones, the hormones can reach the other hemisphere.. This is a means of communication which, although it is less rapid than neuronal communication, is important for mental states. The spinal fluid, of course, is also shared by the two hemispheres and this may be even more important than their common vasculature, a point progressively more important as the number of known cerebral peptides continues to increase.

Another noncallosal means of interhemispheric communication is crosscueing. "Crosscueing" means that one hemisphere initiates a bodily behavior which can provide information to the other hemisphere. One method is "verbal crosscueing": when the left hemisphere names out loud an object being felt in the right hand, this often is followed by correct retrieval by the left hand of that test object. This is in contrast to the customary failure of crossretrieval when the splitbrain human attempts the same task while remaining silent. More intriguing than verbal crosscueing by the left hemisphere of the right is the reverse, nonverbal crosscueing by the right. Nonverbal crosscueing by the right hemisphere takes various forms, as described next.

4. SOME EXAMPLES OF NONVERBAL CROSSCUEING (AND OF ITS FAILURE) IN THE HUMAN WITH COMPLETE CEREBRAL COMMISSUROTOMY.

Systematic investigation of nonverbal crosscueing is as yet scanty. But nonverbal crosscueing has been observed by many people on numerous occasions.. As will be readily apparent. any of these "tricks" of crosscueing could be employed by the normal as well as by the commissurotomized individual.

4.1. L.B. was tested 2 1/2 years after surgery. During the course of the examination, several objects were placed in his left hand while he was blindfolded, and he was unable to name them.

When a paper clip was placed in his left hand he was completely at a loss. He then reached over to put it into his right hand and immediately and correctly named it. (Here, transfer from left to right hand made suffcient information available to the speaking left hemisphere.)

When a pipe was placed in his left hand the pipe was turned around and manipulated in various ways and ended up in a correct position. The pipe was then put up to (but not into) his mouth and he then said, "Oh, that's a cigarette." He then was told to feel it with his right hand and immediately after doing so he said, "No, that's a pipe."

When a pair of glasses was put into his left hand, the hand turned it about and then opened the bows and started to put it on his head at which point his arm was stopped and he was asked to try to name it without trying to do anything further. He was unable to name it. He then flipped the bows closed which made an audible click. He then quickly said, "Are those your glasses?" (Here, the movement by his left hand provided an auditory cue to the left hemisphere.)

A folded handkerchief was put into his left hand which squeezed it, turned it around, and then smoothly reached backward and slipped it into his left hip pocket. At this point he said "Oh, sure, that's a comb." When told he should feel it with his right hand, he did and then shook his head with a chagrined look and said, "A handkerchief." (Here, the lefthand behavior was insufficient to permit accurate identifcation by the left hemisphere.)

4.2. Another example of crosscueing by L.B.

Stuart Butler described (personal communication) a situation in which he and Ulf Norsell were trying to determine if L.B.'s right hemisphere could talk. The idea was to restrict the input to the right hemisphere by using left halffield projection or by putting objects in the left hand which are not apt to provide ipsilateral input, i.e., objects which don't have sharp points, which don't offer an obvious temperature clue, etc. So Butler and Norsell (45) were using a wooden sphere, a wooden cube without any sharp edges, and a wooden pyramid. The question was, could L.B. indicate above chance level whether he had in his left hand the sphere, the cube, or the pyramid? Surprisingly, he was identifying them most of the time! He had his left hand underneath a screen so he couldn't see what was in his hand, and they couldn't understand how he could recognize the objects so reliably. They at first thought it might be some kind of subcortical communication. Then they noticed that he was looking around the room. When they put the sphere in his left hand he would look at the wall clock and he'd say, "It's the round one." When they put the cube in his hand he would look at the door and he'd say, "That's the square one." When they put the pyramid in his Ieft hand he'd look up at the ceiling and pause a minute and then say, "It must be that triangular shape." When Butler blindfolded him, instead of letting him look around the room, his verbal reports of what was in his left hand fell to chance.

4.3. Using chimeric stimuli, Levy et al. asked the subjects to point to the picture that matched what had just been presented; they observed that either hand could indicate a picture selected by gaze. L.B. was noted on several occasions to use scanning of surroundings and gaze fxation to cue his left hemisphere (46).

4.4. Crosscueing by C.C., as observed 8 years postop (13 October 1973).

With respect to anomia, this patient's performance was somewhat variable; it was remarkable, considering his low IQ (below 80 both pre and postop) how he could sometimes identify objects in the left hand on the basis of quite minimal cues. For example, when a pencil was placed in his left hand he was rather slow and deliberate about answering, turning the pencil over in his hand and pressing with the ball of his index finger first on the eraser end and then on the pointed end, and then saying, "It's a pencil."

When a dime was placed in his left hand he rolled it around, rubbed it and said, "It's a penny." He then said "Well, isn't it?" When told that he was not correct he said, '.Maybe it's a dime."

When a rubber band was put into his left hand he first shook his head. However, after rolling it around and then maneuvering it into a position in which it looped over his thumb and opposing fingers so that there was some springy resistance to partial opening of the hand, he said, "a rubber band."

When a pair of glasses was put in his left hand his fingers rubbed the lenses, closed one bow with a snap, and he then said, "Are these glasses?" When he was told that he should try again, he said, "Well, I don't know." When this same obJect was then placed into his right hand he immediately said, "Yes, it's a pair of glasses."

When a pipe (still warm) was placed in his left hand he said, "Is it a cigar?". When he was told that this was incorrect, he shook his head and said, "I don't know." When it was placed in his right hand he immediately said, "It is a pipe."

When a small paper clip was put into his left hand, he turned it over several times and finally said, "I don't know." When it was placed into his right hand he immediately said, "It's a clip." He was then asked "What kind of clip?"; and he replied, "To keep papers together."

A small pine cone was placed into the left hand and he was at a total loss to say what it was. When this same object was placed into the right hand he said, "It's one of those-I don't know what you call it-it comes from a tree." When he was allowed to see it, he said, "Yes it is from a tree. What do you call it?" When he was told, "pine cone," he said, "Oh yes, that's what it is."

From the above, it is clear that C.C. readily recognized and named objects in the right hand, except with the last object (the pine cone) which he could no more readily name in full vision than he could when palpating it in his right hand. In contrast, objects in the left hand were much less readily identified either by verbal description or by naming; any success in this regard apparently depended upon partial information which sometimes led to the correct answer and sometimes to an answer which was related but not correct. The use of pain, from a pencil point, is a cue which I have observed in four different complete commissurotomy patients. Identification of the rubber band apparently involved some proprioceptive information. The glasses provided a definite auditory clue, and possibly a slight temperature clue. The identifcation of a pipe as "cigar" is particularly interesting-it is possibly ascribable to an olfactory cue. Alternatively, his speaking left hemisphere may have received some sort of sparse categorical information either from the left hand via uncrossed pathways, or possibly even from the other hemisphere through internal pathways. In any event, it is quite instructive how an individual with a verbal I.Q. under 70 ( 65 at this time) can use minimal sensory clues to identify objects which are familiar to him.

4.5. Crosscueing (mainly unsuccessful) by R.M., when tested 4 months postop (13 July 1966).

He had the usual anomia in the left hand with his eyes covered. When a pencil was placed in his left hand he held it appropriately but could not name it. It was then put into his right hand and he said, "a pencil." When a watch was put in his left hand, he said it was a "pencil" even when, with his left hand, he was holding the watch up to his left ear. A paper clip was put in his hand and he could not tell what it was; but when he put it in his right hand he immediately identified it. A pipe in the left hand was put into his mouth in an appropriate way; but it was called a "pencil" even after the bit was between his teeth. When an ashtray was put into his left hand he struck the table with it; it made a distinctive sound and he immediately told me what it was. When a pair of glasses was put in his hand, he could not name what he was holding until he tried to put them on. A handkerchief was put in his left hand; his left hand immediately put it into his Left hip pocket [in the same maneuver as used by LB] but he could not say what it was. When he felt it with his right hand he immediately named it correctly.

4.6. Less usual kinds of crosscueing.

Some "tricks" of crosscueing are quite common across patients (in spite of their only rarely if ever meeting one another). Other "tricks" are more obviously idiosyncratic. Moreover, some "tricks" are easy to recognize, whereas others are only doubtfully identifiable. Four patients (A.A., L.B., C.C. And N.W.) have been personally observed by me, each on several occasions, to turn a pencil in the left hand so the sharp point could be pressed into the distal volar surface of a finger. And I have personally observed five (A.A., L.B., C.C., R.Y., and R.M.) loop a rubber band over several fingers to explore its elasticity. (Here, there is probably some ipsilateral proprioceptive information from the left hand to the left hemisphere. )

On the other hand, the following are "tricks" observed only on single occasions:

(a) On one occasion N.G. was being tested, in full view in the normal way, on the Benton-Van Allen faces test. Part way through the test she changed her manner of responding. Instead of pointing quickly with her hand (to one of six choices), she would pause before her choices and move her hand only after a motion of her head which resulted in pointing of her chin toward the choice subsequently pointed out by hand. When this was recognized and she was asked not to move her head she resumed the undelayed pointing with her right hand.

(b) (Courtesy of E. Zaidel) On one occasion L.B. was being tested by presenting Peabody pictures to one halffield or the other. After a picture of a tree appeared in the left halffield, his hands formed a triangle, and he then said,"teepee".

5. OTHER SOURCES OF "MENTAL DUALITY"

In addition to the noncommissural communications between the two hemispheres in the splitbrain, there is another issue needing clarification. That is, there are clearly other sources of duality or of ambivalence besides having two hemispheres.

We are familiar with the distinction between a pyramidal and extrapyramidal system. We know that monkeys can function remarkably well after the pyramids have been transected bilaterally. All movements are commanded from the spinal cord, but what the spinal cord does can be directed from above in more than one way, and these directives might well be both concurrent and conflicting. So one can see here some opportunity for ambivalence or conflict which requires only one hemisphere.

There appear to be at least two kinds of memory (in each hemisphere) which are anatomically specifiable: one is procedural memory probably subserved by the basal ganglia; a second is declarative memory (episodic and semantic) dependent on medial temporal structures. These facts suggest that a dissociation between episodic memory and procedural memory (I.e. Between explicit and implicit memories) might require only one hemisphere.

There are many kinds of duality which can be checked, and should be checked, in hemispherectomized individuals because, if one saw that kind of duality in them, then there would be no need to suppose that it depends on the duality of the hemispheres. Unfortunately, very little has been done along this line. Indeed, in view of the widespread availability of hemispherectomy patients, it's astounding that this opportunity is not Ibeing pursued to settle convincingly the relevance of hemispheric duality for various kinds of "mental duality" (i.e. dissociations between what the subject says he knows and what we believe he knows on the basis of nonverbal behaviour).

Having argued that considerable inter- hemispheric integration occurs absent the corpus callosum ,because of a host of other unifying mechanisms, behavioral as well neuronal and hormonal, I will move on now to the complementary claim, that even with the commissures intact, there is a significant degree of hemispheric independence. Before reviewing experimental data, I will first offer a quantitative argument for a significant independence of hemispheric function.

6. A QUANTITATIVE ANATOMY ARGUMENT FOR SIGNIFICANT HEMISPHERIC INDEPENDENCE WITH NEOCOMMISSURES INTACT

According to both Leake (48) and Garrison (49), the first time this kind of argument was used in medicine was by William Harvey in 1628. Harvey measured the volume of the heart and counted the heart rate. He asserted that the amount of blood that is pumped by the heart in any given time (say in a day or an hour), could not possibly be produced in such sizable amounts by the lungs. Therefore, the blood must circulate. He did not demonstrate the circulation; he simply concluded that it must be so, on this quantitative argument. And it was more than 30 years later (1661) that Marcello Malpighi actually demonstrated the capillaries with the microscope.

An argument from quantitative anatomy in the present case is roughly as follows: There are at least 200 million nerve fibers in the corpus callosum (50). By contrast, within each hemisphere there are at least 10 billion nerve cells,each with at least 100 connections (and many nerve cells, of course, have thousands of connections). This means that the number of connections between the hemispheres is less than 1% of the connections within each hemisphere. Hence,that the information generated in one hemisphere can be immediately transferred ormade available to the other is hardly credible.

We might suppose that information accumulated by one hemisphere during waking could be transferred later, during sleep; but callosal activity decreases markedly during EEG synchronization and/or REM sleep (52). In any event, immediate sensory information in one hemisphere seems fully available to the other only with respect to the midline, since for primary sensory cortices it is the areas for midline representation which have heavy callosal projections (53-56). Furthermore, as Marzi recently wrote:

"Since there is a clear trend . . . for a progressive increase in the wealth of interhemispheric connections as one moves from the primary visual cortex to extraoccipital visual areas, a reasonable assumption is that the main burden of interhemispheric exchange of visual information concerns late processing stages". (57).

According to evidence to be considered next,higher order information (as accompanies learning) is not necessarily shared or hemispherically equivalent even when the cerebrum is anatomically intact.

7. EXPERIMENTAL EVIDENCE FOR SIGNIFICANT HEMISPHERIC INDEPENDENCE WITH RESPECT TO LEARNING.

The experimental evidence to be considered here is of different kinds, but it has mainly to do with incomplete interhemispheric transfer. Whereas we usually emphasize the lack of interhemispheric transfer in the split condition, the many authors who have re ported the splitbrain animal experiments have sometimes alluded, albeit briefly, to the fact that in the intact case there is often (especially in the unsophisticated animal) a lack of transfer. In general,if one trains a monkey to do something with one hand, the monkey often doesn't straight away do it as well with the other hand. After a while the monkey transfers better; that is, the lack of transfer typically goes away, unlike the anatomically split condition where the lack of transfer persists.

A previous review (4) described some evidence for lack of interhemispheric transfer of information in the intact brain including the cat experiments of Myers (58). Also mentioned there was a lack of transfer of spatial form discrimination from one hand to another during early testing of a difficult task in monkeys (59). Butler and Francis (60) reinvestigated this often evanescent phenomenon. They arranged for the subjects (baboons) to reach through a tube to manipulate stimuli at the other end, thus restricting movement of proximal joints. The animals learned to rotate the stimuli (knobs)to get food, in either a clockwise or counterclockwise direction, depending on the shape of the knob.. After the subjects learned the task, the other hand was tested. Two different problems (A and B) were used (three animals learned A first and three learned B first). On the first problem, whether it was A or B, the authors found:

". . . no animal was able to perform without further training the discrimination learned with the other hand; in all cases the animals required extensive training with the second hand before regaining the proficiency reached with the first one.".

On the second problem (whether B or A), rather than performing at chance as on the first problem, there was often an abundance of errors because of performance in a mirror image mode. These results could be interpreted as showing the development of a learning set, that is, the animals may actually learn to transfer information, with repeated testing. A similar tendency to acquire transfer was observed by Berlucchi et al (61). A lack of intermanual transfer need not rule out interhemispheric communication; it is possible that in some cases information is transferred but that it is not initially used by the receiving hemisphere..

Much earlier, in 1949, Sperry and Clark reported related findings in gobies. (lt is worth recalling that the splitbrain story started with the question of interocular transfer, i.e., whether learning transferred from one eye to the other.) In this paper Sperry and Clark wrote

". . . the results seem to support the conclusion that the brain organization of this teleost fish permits interocular transfer. but at the same time the neural mechanisms involved are not so well developed that good transfer is automatically assured in all instances ".(62).

In fact, clearcut savings were observed in only 5 of 16 cases. Lesser degrees of transfer were found in 4 whereas transfer was essentially nil in 7. The l l without good transfer had not simply "forgotten" the task, because retesting of the originally trained hemisphere showed approximately the previous (trained) level of correct responses. Then the untrained eye was again tested in 2 cases, and excellent transfer was found in I of them.

In considering the above experiments with fish, we recall that the chiasmal crossing in these animals is essentially complete, so that training one eye is tantamount to training one hemisphere directly and the other only indirectly (if at all). A nearly complete chiasmal crossing is also present in the rat and rabbit, which we consider next.

Russell and Morgan (63) directed the input in the white rat to one hemisphere by restricting the input to one eye. (The white rat's chiasm is almost totally crossed, the wild rat has a little less crossing, whereas in the human or monkey it is approximately 50%.) Then they tested the other eye. What they concluded was that, in rats, a failure of interocular transfer need not require an anatomical disconnection. On the contrary, as they said, ". . . these results suggest that under certain conditions an absence of interhemispheric communication is a characteristic of the intact brain." [It is worth emphasizing here that transfer should be better in a normal rat than in a human because the two thalami in a rat are closely coupled by impressive midline thalamic nuclei whereas in the human the two thalami are commonly not directly connected and the midline nuclei are negligible.]

Using rabbits, Van Hof repeatedly found "lateralization of the memory trace"; but he cautiously concluded that "it would be premature to regard the rabbit as a 'natural splitbrain preparation." (64). The rabbit has a corpus callosum, of course; but the engram may not transfer if one sets up the learning situation in the way used by Van Hof.

Related results have been found in pigeons by a number of investigators, beginning with Beritov and Chichinadze (65) and recently considered in depth by Graves and Goodale (66, 67); they concluded that transfer-or lack of it-depended upon the type of problem. Also apparently depending upon the experimental conditions, either unilaterality or bilaterality of the engram has been found in newborn chicks (68-71).

In this context (monocular training and total chiasmal crossing) it is instructive to consider a case which reinforces a previous point of this essay: that we could mistakenly overemphasize the importance of the corpus callosum for interhemispheric communication. Marzi et al. (72) used Siamese cats, whose optic pathways fully cross, permitting restriction of input to one hemisphere. They found a deficiency of interocular transfer on some more difficult problems, as compared with cats having normal optic chiasmata. The Siamese cats did show essentially complete transfer on simpler problems: and they did so even with section of the neocommissures. The authors suggest a role for subcortical connections, as these might be more effective in Siamese cats in compensation for their chiasmal peculiarities.

For visual problems, in the experience of most investigators, monkeys usually show transfer so long as the splenium is intact (73-77). But even with all commissures intact, lack of transfer has occasionally been evident as a learning deficit on testing of the second hemisphere.

What is important for the purpose of this paper is that there are certain learning situations which give rise to certain kinds of information which do not readily transfer

The likelihood that information will be transferred apparently depends upon (among other things) the functional ability of the hemisphere receiving the transfer, particularly if the ability has been reduced by a lesion. Zeki did an important experiment along this line. After making partial lesions in the parastriate cortex, he taught chiasm-sectioned monkeys to do a problem with one eye or the other. When the eye on the side of the lesion was trained first, there was very good transfer. But when the eye on the unlesioned side was trained first, the second eye did not show transfer. He suggested, because the unlesioned side learned with fewer trials, that the interocular transfer in the original circumstance was due to overtraining; a unidirectionality of transfer (for the problems he used) was not found if lesions were placed on both sides (78).

In addition to lack of interocular transfer, another manifestation of hemispheric independence is an inability to crosscompare objects presented to separate eyes. Voneida and Robinson (79) found this in two chiasmsectioned cats. These authors were also able to train contradictory habits to separate eyes when the rostral onethird of the callosum was intact and there was still some interocular transfer. Other examples of unilateral engrams include a conditioned salivary response with surgically split tongue of the dog (80) and a conditioned auditory response in cats (81).

7. RESULTS WITH CORTICAL SPREADING DEPRESSION (CSD), INCLUDING SOME METHODOLOGICAL CONSIDERATIONS

Next we consider the cortical spreading depression (CSD) experiments. If one puts a drop or two of concentrated potassium chloride solution on the cortex of a hemisphere, especially a smooth brain like that of a rat, waves of depression spread over the cortex so that the entire hemisphere becomes temporarily inactivated. Meanwhile, one can train the other hemisphere. Then after the potassium chloride has been washed away, the rat has two good hemispheres, but only one which has learned the task. If one then inactivates the trained hemisphere with CSD, the other hemisphere doesn't know the problem. So even though the animal had two hemispheres functioning normally for a while, the learning didn't transfer. Bures and colleagues have reviewed this subject in great detail(82). The main point is that CSD provides an independent kind of information leading to the important conclusion; that there may be lack of interhemispherictransfer with commissures intact.

Experiments using CSD have sometimes been discounted (31, 83) on the objection that CSD experiments have serious methodologic deficiencies. But, as we shall see, this objection is only narrowly applicable. Before examining this objection, we first recall that every technique has its methodologic difficulties and ambiguities; we can consider, as a notable example, the various techniques for cerebral localization.

Cerebral localization (CL) in cases of brain tumor has been extraordinarily productive: many of Hughlings Jacksons' notable conclusions regarding CL were based on tumor cases (12). Yet it is generally recognized that tumors can be misleading as a result of pressure effects~ or because of ongoing diaschisis in the tumors of rapid growth. The phenomenon of tumor ''momentum" is well known to neurologists. Indeed, these and related phenomena gave rise to an entire literature on "false localization."

Because of the problems associated with CL in tumor cases, it has often been asserted that CL should be based not on tumor cases, but on cases of fixed deficit from thrombotic stroke, preferably cases eventually coming to autopsy, although in recent years CAT scan and MRI localization have significantly complemented autopsy findings. Using stroke cases for CL has a long and illustrious history. But clinicians are aware that the putatively critical infarct is usually accompanied by other infarcts, previously occurring and supposedly "silent," and that these other infarcts contribute in all likelihood to the behavioral deficit consequent to the "critical" infarct. (A wellknown example is that when a retrorolandic lesion in the right hemisphere produces a longlasting deficit in face recognition (prosopagnosia), it typically has been preceded by a lefthemisphere lesion, often asymptomatic.) Moreover, considerable ambiguity can be present in cases of single thrombotic infarcts, because these usually occur in a brain already affected by arteriosclerosis, with lessened compensatory reserve or even, as in many cases, with overt manifestations before the stroke.

The ambiguities of CL in older patients with strokes can be avoided by studying youths wounded in war. Luria's book on traumatic aphasia, which really made his reputation, was based on Russian soldiers from WWII (84). Marie and Foix described French soldiers from WWI (85). Russell and Espir (86), as well as Newcombe (87), studied British soldiers from WWII. Some very important studies have been based on missile wounds in young people. Of course, these are not free of problems either. It's not altogether clear where the boundaries of a missile injury are. What may be even more problematic is that the missile can cause vascular derangement, for example, spasm in a blood vessel, which then results in focal damage to brain tissue at a considerable distance from the missile track.

Perhaps all of this tumor, stroke, and missile evidence should be discounted, and we should depend on cases of surgical removal where the boundaries of the ablation have been precisely described. The work at the Montreal Neurological Institute by Brenda Milner (88, 89) and her students and colleagues has utilized just that sort of material. However, we are not always persuaded of the total reliability of the surgeon's reports, especially now that we have experience with postoperative CT scans. Besides, no matter how precise or reliable the surgical removal, these studies concern cases of longstanding epilepsy, so that there are ambiguities attributable to the compensatory changes that occurred subsequent to the epileptogenic lesions.

The same sort of problem comes up with hemispherectomy, which is often as anatomically clean as any removal can be. In such cases anything in the form of behavior is attributable to the residual hemisphere. But most hemispherectomies have been done for infantile hemiplegia which means that people who have had those operations have had a long disease history with opportunity for a lot of compensatory change. Even when hemispherectomy has been done in adults of normal development, and that's quite rare, there is some delay between the operation and behavioral testing during which compensation could occur. In fact, compensation has been documented with repeated testing (90).

The confounding effects of compensatory change can be avoided with the "temporary hemispherectomy" of the carotid amytal (Wada) test. But here we have other problems including that the posterior cerebral artery of the injected side often does not feed from the carotid artery. Moreover, carotid branches to the brain stem can seriously affect the behavioral results of the injection.

Does this mean that conclusions from all of the foregoing types of material should be dismissed? Certainly not! The point is that every patient population and every experimental method has its problems. This is also true of the EEG approach to CL, the use of blood flow measures, the PET scanner, and stimulation mapping. as well as the split brain.

The lesson here is twofold. First, one does not summarily dismiss data having interpretive complications; one takes the time to become familiar with the qualifications and ambiguities, and then relies upon the data with appropriate reservations.

Second, a conclusion indicated by any method or material must remain tentative until confrmed by other methods having different methodologic problems. When several different approaches point to the same conclusion, we can have increased confidence in the result. It is convergence upon a common conclusion,through a variety of different approaches, which gives us a near certainty in our conviction of complementary hemispheric specialization in the human. Similarly, it is the convergence of a variety of evidence which leaves us in little doubt aboutthe duality of the brain. One kind of evidence for cerebral duality comes from cortical spreading depression (CSD) and we now turn to a closer appraisal of the supposed deficiencies of this approach.

Gazzaniga and LeDoux (31) and Denenberg (83) dismissed CSD evidence largely on the basis of a review by Petrinovich (91). What concerns us in this paper is whether the methodologic difficulties pointed out by Petrinovich weaken the conclusion described above, namely that an engram originally engendered in one hemisphere (while the other was subjected to CSD) does not automatically transfer to the other hemisphere when both are returned to a normal state. In general, for the untrained hemisphere to exhibit the learning on its own, when the orginally trained hemisphere is depressed, a number of training trials are necessary with both hemispheres undepressed. That is, substantial transfer commonly requires interdepression training (IDT). What we find on careful examination of the Petrinovich review is not an objection to this conclusion. Rather, what we find is a critique of other inferences from CSD studies, in particular the belief that the engram is always completely lateralized, or that when learning is lateralized, that IDT produces transfer with a specifc number of trials. . .

At first glance, Petrinovich's review appears to be summarized in his closing clause: ". . . it is concluded that the technique (CSD) is of questionable utility in elucidating with any precision mechanisms involved in the formation and transfer of memory traces." But it is the ''precision," not the fact of transfer, which is the object of his critique. He specifically singles out for censure the conclusion of Albert that it takes only 3 min for the trained hemisphere to transmit the memory (during IDT) but it requires 2 h for the transmitted memory to consolidate in the untrained hemisphere. The greater part of his review is devoted to the sources of error in three articles by Albert (in Neuropsychologia in 1966), and two papers by Mayes (in Behavioral Biology in 1973) whose findings were congruent with the findings of Albert. The errors included running animals from light to dark, assuming that washingoff KCI quickly restores cortical normality. discarding the results from recalcitrant animals, and the use of inappropriate statistical procedures. Petrinovich does not deny that it is possible to train one hemisphere while keeping the other one untrained with CSD, or that the memory trace or engram subsequently can remain largely lateralized even with both hemispheres functioning. As Petrinovich says in the opening sentence of his discussion, "it is generally agreed that IDT produces transfer, where transfer is defined as the savings observed when the initially depressed cortex is trained undepressed following IDT. "

One of those [Denenberg (83)] who discounted CSD data relied not only on Petrinovich but also on the review by Gaston (92). But her criticisms are concerned solely with the question of taste aversion, that is, whether CSD experiments can prove the participation of cerebral cortex in a rat learning to avoid certain tastes, learning which possibly can occur without cortex at all. Nowhere does she deny that an originally unilateral engram can remain lateralized in the posttraining, undepressed state.

In conclusion, CSD experiments have provided some very clear examples of the failure of intact commissures to effect a synchronization (in the sense of equivalence of information content) of the two cerebral hemispheres. [{ This story also shows how hasty reading can lead to misreading and malusage of the literature}].

8. HEMISPHERIC INDEPENDENCE IN THE INTACT HUMAN

We consider next the intact human. In the young child, interhemispheric communication might be deficient since myelination of callosal fibers takes 10 or more years to reach completion (93). In fact, deficits in tactile crossmatching have been found (94). But in the adult, lack of interhemispheric communication has been less directly demonstrable.

A dramatic result was reported by Risse and Gazzaniga in 1978 (95). Eight righthanded candidates for neurosurgical intervention were having carotid amytal testing. After injection of the left carotid, the patient having become right hemiplegic and aphasic, the examiners presented, in the midst of other testing, some objects to the view of the patient. When the patient subsequently was moving well and talking well, they asked, "What were the objects we showed you?" Six patients replied, "I don't remember that you showed me anything." And they said, "Here are some objects;point to one"; and the patients pointed to the correct test items. What the authors concluded was that the nonverbal engram can remain inaccessible to linguistic report. It's possible that this result reflects two "kinds" of memory in one hemisphere, since recognition is different from (and easier than) recall with verbal report (89). Moreover, we still await replication, which has not been found by others who were looking for it (96). [{A review of Wadatesting in the 1990s is needed here}].

A lack of interhemispheric transfer in the intact human has been demonstrated in other ways. It seems reasonable to suppose, as did Butler and Francis (60), that manual manipulations will show less transfer to the other side, the more distal (in the extremity) are the crucial aspects. This could be expected in humans as well as monkeys, hence it is no surprise that it is for fine digital manipulations or detections that savings seem least. A good example is the learning of Braillelike patterns by sighted adults new to the task. Although there are usually some savings, it requires further training with the second hand to reach the same level of proficiency as with the first hand. Indeed, when going from left to right in right handers, just as much training may be needed for thesecond hand (Wagner, 1977, reproduced in Harris, 1980) (97). The amount of transfer probably depends upon what the second hand is doing while the first is being trained. Hicks et al. (98) had subjects learn a sequence of key presses on a typewriter with one hand while the second hand was either resting or grasping a table leg; partial transfer occurred to the resting hand, but not to the hand occupied with grasping during the original training.

An ingenious study using dichotic listening to digits and tones was reported by Goodglass and Calderon (99), and their conclusions were supported in another dichotic study by Sidtis and Bryden (100). This is not even a situation in whichthe material is clearly directed to one hemisphere or the other. These results can be summarized by quoting the discussion by Bradshaw and Nettleton (101) their excellent review, Human Cerebral Asymmetry, in which they say: "These results show that the two hemispheres could concurrently and independently process that component of a complex stimulus for which each is dominant."

Next we come to notable experiments by Landis and coworkers (102). Their purpose, in the beginning at any rate, was to ask: Is there a right hemisphere superiority for the recognition of facial expression? We know that the right hemisphere is predominant, although not exclusive by any means, for the recognition of individual faces. The question naturally arises, is the right hemisphere also predominant for the recognition of facial expression? They did a visual halffield test as follows: They presented centrally a schematic outline of a familiar object, and in one or the other halffield a photograph of the same thing or of something else and asked, "Are these the same?" For example, a schematic drawing of a corkscrew appeared in the middle at the point of fixation along with a photograph in either the left or right halffield; if it was a photograph of a corkscrew the subject would push the button. If it was different, no reaction was necessary. And they measured the reaction time. They did the same procedure with facial expressions. A schematic diagram of a front view face appears in the middle, along with a photograph, in either the right or left halffield, of a face in profile with an expression (happy, sad, etc.). Is it the same expression or not? It turned out that the reaction times are less in the left half field for the facial expressions whereas the reaction times for objects like the corkscrew are less in the right halffield. Now this is not altogether surprising; it's what most of us would expect on the basis of earlier evidence for right dominance for faces. We might also expect that people would be more accurate in recognizing the similarity of objects than in recognizing identity of facial expressions; indeed, the responses were more accurate(96% correct) for the objects than for the expressions (76% correct). [{Right hemisphere predominance for recognition of facial expressions seems not to be true for lipreading, presumably because the facial recognition aspect is overridden by linguistic demands (103).}]

Landis et al. also noticed that the subjects showed very different levels of awareness of their decisions, often having second thoughts. That is, erroneous object matchings were often followed by rapid recognition of the errors; and accurate object matchings were confirmed. By contrast, when matching expressions the subjects repeatedly expressed the sense of having made an error when, in fact, the manual response had been accurate. In other words, the subject made many accurate decisions on facial expressions without correctly monitoring their own behavior. "However," the authors wrote, ''this is an incidental observation which needs to be explored systematically" (102).

There subsequently appeared a paper by Landis with Graves and Goodglass; the title of the paper asked, "A SplitBrain Phenomenon in Normal Subjects?" (104). The same stimuli were used, but they modified the method of presentation. First, the photographs of objects and faces were presented randomly in one or the other halffield. Also, they made the presentation time sufficiently shorter for the objects so the subjects were wrong about 25~% of the time with both the expressions and the objects. After responding, the subject was expected to report, "whenever you think you made a mistake." The principal point here concerns the individual's monitoring of his own behavior. In this second experiment the authors were not looking at the reaction times but wanted to systematically inquire after the subjects’the second thoughts. They found that the monitoring with the object matching was very good, whereas the monitoring for the expression matching was at chance level-the verbalized corrections did not reflect the person's abilities to make accurate decisions about expressions. That is to say, when the behavior was presumably controlled by the right hemisphere, second thoughts were not reliable. In the author's words: "The result for the emotional expression matching task resembles results obtained with splitbrain patients."

Hemispheric capabilities have been studied in three principal ways: from studies of patients with lateralized lesions, from studies using lateralized input (as in the halffield studies just described), and from studies using lateralized readout (including both radiographic and electrical methods). The final bit of evidence I will present concerns the last, namely, lateralized readout in the form of smatosensory(tactile)evoked potentials. The experiment was done by John Desmedt (105, 106). He used as a stimulus the flat end of a cylindrical (dowelshaped) rod. The end of the dowel briefly and repeatedly touched the ball of a palpating finger. If the end of the dowel had a ridge on it, the subject's task was to determine the orientation of the ridge, (eastwest or northsouth). If there was no ridge, the subject did not have to make any spatial orientation judgements. In this experiment, each time the dowel touches the end of the finger, it evokes a potential recorded from the scalp. And each time, the subject has to decide whether there's a ridge or not,and if so, how it is oriented.

The end surface of the dowel, when it was smooth, elicited a potential which was the same overeach of the two hemispheres. However, when the end came up with a ridge, there was a different potential evoked over the right hemisphere, whereas for the left hemisphere the fact that there was a ridge on the end caused no change. Let me translate for you the author's conclusion:

"These observations are notable because they show the localization in the right hemisphere of cortical mechanisms put into play by the perception of spatial orientation in normal subjects. They show that the capacities of the right hemisphere inferred from neurologic studies of individuals with lateralized lesions do not result from a liberation from control of the left hemisphere. These findings are therefore in favor of the view that there is not a unilateral dominance by the left hemisphere and they are in favor of the notion that each of the two hemispheres possesses different capacities and that the capacities of the right hemisphere can be manifested [independently]even when the commissural functions of the corpus callosum are intact (105).

Since the foregoing was true with each hand, it would appear to involve, in the righthand testing, a callosal relay rather than direct access (as distinguished by Zaidel in 1983 (107)). That is, it seems a case of unihemispheric function (with lateralized electrical sign) which is both preceded (for relay of input) and followed (for verbal report) by interhemispheric communications.

It is understandable that someone would say, (as some do from time to time) that two hemispheres acting together are better than one, especially if we suppose that they are interacting in a mutually supportive manner (108). That two should be better than one, and moreover that two should incline to act together rather than separately, seems to be in accord not only with our usual intuition but also with some objective, experimental facts. Indeed, we have previously emphasized the crucial importance of hemispheric interaction (4). Of course, in the human the two hemispheres are different (101, 109112). This might incline sometimes to interference rather than simple reinforcement, something to be determined, in al1 likelihood, for each particular circumstance.

That two brains should incline to work together and that when they do they are better than one does not, however, demonstrate that there was but one brain to begin with. In fact, it seems actually to be a sort of backhanded argument for the twobrain view, an argument of interest even if a bit redundant for those who have already found the forehanded evidence more persuasive.

9. BOTH ONEBRAIN AND TWOBRAINS: A CASEOF COMPLEMENTARITY?

Descriptions of human behavior in terms of "one brain" or of "the mind" (singular) have been with us for a very long time. Indeed, they have been in use for so long, and have been so often helpful that we can rest assured that such usage wil1 continue.

On the other hand, there are circumstances for which the "one brain" interpretation has not sufficed. More and more people find it not only convenient, but fruitful to attribute to each hemisphere the qualities of a mind (25). The number of such persons is growing. Moreover, since in science as elsewhere one finds what one looks for, we can expect that observations consonant with or even demanding of such "two brain" usage wil1 continually accrue. Those who find such usage uncongenial could find themselves in a progressively shrinking minority.

It seems to be necessary to change from one usage to another, depending on the circumstances. We might then ask if this is merely a matter of verbal usage. Is it possible that our describing the bram as sometimes single and other tmes double is related to our vacillation between mentalistic and physiologic explanations of behavior? Most of us do resort to both of these approaches, emphasizing one or the other as it seems to suit our needs.

How can we hold simultaneously two seemingly incompatible (and occasionally conflicting) descriptions? Possibly relevant here is a wellknown precedent in quantum physics, whose practitioners learned to hold both the wave and particle views simultaneously. Einstein expressed it in 1924 as follows: "We now have two theories of light, both indispensable, but it must be admitted, without any logical connection between them, despite 20 years of colossal effort by theoretical physicists" (113). A halfcentury later Holton wrote, "One cannot construct an experiment which simultaneously exhibits the wave and particle aspects of atomic matter. A particular experiment will always show only one view or [the other] " (113).

If the mentalism/physicalism pair persists not simply as verbal usage but because no experiment can decide between them, we might consider that they make up what Niels Bohr called a "complementarity" (114). As pointed out by Holton, Bohr felt that the concept of complementarity was applicable not only to the wave/particle contrast but that it could be particularly helpful in psychology (113). The idea that quantum theoretical concepts could help us understand brain states has often been argued by D. O. Walter (115, 116). What is complementarity ?

By the term "complementarity" we refer to the existence of two conflicting interpretations, such that either interpretation provides a useful but incomplete description. As a result, both interpretations must be utilized (usually by alternating or oscillating between them) in order to have the fullest possible account. We do not say that a complementarity exists whenever we have two conflicting explanations of something. Indeed, the overwhelming majority of such pairs are not complementary. It is a prominent aspect of scientific endeavor to devise experiments which will help us decide between one or the other explanation, to choose what we call the "correct" interpretation. In order for a mutually conflicting pair to be considered a "complementarity," at least two conditions must be fulfilled:

(i) After numerous attempts to formulate a decisive test, either in actual practice or by thought experiment, it becomes accepted by workers in the field that no such experiment is possible [Feynman (117)]. Since this conclusion is necessarily dependent upon the somewhat dubious belief that current workers in a fleld are as knowledgeable as they will ever be, a certain reserve inevitably accompanies any belief that no such crucial experiment will be forthcoming. Such doubts as continue to exist in spite of the lack of a decisive experiment can to some extent be allayed by adding a compelling, rational argument, such as the following:

(ii) It is realized that the two conflicting views possess certain essential properties in common, whatever may be their superficial differences, so that no crucial "distinguishing test" should be expected. Such a criterion for waveparticle complementarity was reached when it was realized that a mathematical formulation of the wave theory (the wave equation of Erwin Schrodinger) was of sufficient scope and power to account for both aspects. Not only was Schrodinger's equation shown to be isomorphic with the matrix algebra of Heisenberg, but it could be used to predict discontinuous intervals (i.e., quantum jumps) in spite of being a differential equation involving only continuous functions of continuous variables [Feynman (117)]. Heisenberg put it as follows: "Although the classical theories of the corpuscular and wave pictures are so entirely different, both physically and mathematically, the quantum theories of the two are identical" (118).

(iii) There is a third, more technical criterion to be discussed below, after some further generalities.

It is their apparent incompatibility when the two views are expressed in ordinary language, but not when expressed in mathematical terms, which is crucial. As Holton wrote: "What Bohr had done in 1927, was to develop a point of view which would allow him to accept both members of the (a, å) couple as valid pictures of nature, accepting the continuitydiscontinuity (or waveparticle) duality as an irreducible fact, instead of attempting to dissolve one member of the pair in the other . .. Bohr asked that physicists accept both a and å athough both would not be found in the same plane of focus at any given time . . . We see once why all parties concerned, both those identified with a and those identified with å, would not easily accept a new thema which saw a basic truth in the existence of a paradox that the others were trying to remove." (113).

It is worth reiterating how important is the difference between the mathematical formalism (in this case, quantum mechanics) and descriptions in natural language. As Rohrlich wrote,

". . . we find that our common ianguage is utterly inadequate for the description of the quantum world and that the mathematical language is much more suitable. [In quantum mechanics an electron] is "just like a particle" or "just like a wave" only in limiting cases depending on the particular experiment." (1l9).

There is a nice way to picture the waveparticle complementarity and its dependence upon our choice of observational method. This was pointed out to me,in conversation, by Professor F. Zachariasen and goes as follows: consider that the event to be observed has a distribution of definite width (we would ordinarily describe the width as a function of the variance). Then, if the resolving power of the observational method were quite wide compared to the width of the observable, one could consider the latter to be negligible; i.e., we would see the observable as a spike with a definite size (height) and a definite location, but no area (i.e., extension) in the ordinary sense. However, if our observational method had a very narrow field of view, we would appear to be confronted (as we move our aperture around) with a standing wave of essentially infinite extent whose height would be different at various locations and whose area we could determine between any two reasonably close locations, but for which the ordinary concept of location would be inapplicable.

If we could formulate a resolution of what we see now as continual vacillation between two different views, what might be the quantitative concomitants? Does it make sense that "mind" resonates between the two states: "single" and "double"? Could it make sense to describe hemispheric interaction in terms of a phase angle whose size represented the amount of hemispheric synchrony (or disparity)? Does the method of testing (for example, psychological testing) actually affect in some degree whether we see the mind as double or single? This sounds at first like something we might accept. What most of us may not yet be prepared to accept is that the same observable will differ from time to time and that this indeterminacy is not an observational deficiency but is, indeed, the real nature of mind.

10. ARGUMENTS AGAINST COMPLEMENTARITY

Abrupt transitions from one state to another (as from a "onebrain state" to a "twobrain state") certainly do not require the assumption of an underlying complementarity; abrupt transitions of water to ice or vapor are familiar examples. Complementarity could be relevant if it were the case (which it is not for water) that a gradual increase in temperature (or other variable) could have on different trials different results, each with a certain probability but none fully determined. Accepting complementarity (i.e., accepting simultaneously two mutually incompatible views) is something most scientists will not do unless it is forced upon them in each specific case by experimental evidence. (No matter how convenient, seemingly rational, or esthetically pleasing it might be!)So we come to the third criterion alluded to above:

(iii) This third criterion is essential to any actual application of quantum mechanics; and it is claimed, by some at least, that this criterion should be satisfied whenever one wishes to introduce the idea of complementarity. The criterion is that when there are two or more routes to arrive at some result, the probability distribution of the result must be predictable on the assumption of superposition of probability amplitudes (expressed in complex numbers) rather than probabilities. This is the case when the probabilities of the alternative routes are not essentially independent. There is a substantial body of opinion, e.g., Putterman (120), that no such interaction should be expected on the macroscopic level. If that is so, not only would the concept of complementarity be inapplicable to the onebrain/twobrain issue, but indeed (Niels Bohr notwithstanding) to almost any problem of brain function.

Although the assumptiom of superposition might be necessary, it is not sufficient. Suppose we found that a behavior having a Gaussian distribution when exhibited by persons with hemispherectomy had, when exhibited by intact individuals, an interferencelike distribution (more or less Wshaped); then we could conclude that there was an interference reminiscent of quantum mechanics. But this would not justify the conclusion that there is a "deep complementarity" in the physicists' sense. This objection can be expressed more generally as follows: although acceptance of contradictory models, using sometimes one and sometimes another, may be necessary in a wide variety of complex situations (indeed, it is typical of real life), one should not consider this to be "complementarity" in the physicists' sense. For physicists, complementarity asserts a fundamental limitation on our ability to describe the basics of nature in ordinary language, whereas in the complex situations of real life the alternating use of contradictory models, and the airing of a diversity of views, merely reflects our current lack of understanding (121).

This is an enlightening argument but one wonders if we can ever feel we know it all about anything. If complementarity or superposition (or any other idea) can be used to throw light on a subject, who is to say that it is not appropriate only because in that subject (unlike physics) we are not yet "fully knowledgeable"?

Even if a quantum theoretic approach were useful, it might not be needed, smce there are now available a number of other methods for dealing in a classical, deterministic way with the problem of more than one stable outcome of shared probabilities. These include for example, variations on the population equation (122). [{ since this was written, there has been cosiderable development of complexity and chsaos theories going well beyond the population equasion{[. Should such approaches be workable, we might still have (without complementarity) the possibility of both singularity and duality of mind at different times. But then, as in the case of a continuum between one mind and two, all of the observable states would be interpretable in terms of just two minds and their combinations, including overlap or intersection in varying degree.

Among neuroscientists there remain many who still find uncongenial the twobrain view. To them we say: find an experimental situation which is explained by the onebrain view and not by the double brain. There are now available data which are inexplicable on the onebrain view; are there also objective data which are incompatible with the twobrain view? If not, considering complementarity may not be necessary, or even reasonable. And we can then look forward to the eventual demise of the doctrine that our two hemispheres make up but one brain.

11.CONCLUDING REMARKS

The splitbrain affords us an anatomically defined circumstance in which a single individual can manifest two minds, each with its own discriminative, mnemonic, and volitional capacities. Suppose that we restored the corpus callosum; would that reduce the complexities? To put it differently, if we added 200 million bridging nerve fibers, would a previously complicated situation (that is, mental duality) become more simple? We might even suspect that the mental numerosity of the intact brain is greater that that of the splitbrain.

If an individual with a hemispherectomy has one mind, and two individuals each with a hemispherectomy have two minds, where between them shall we place people with splitbrains, who are clearly less double than two individuals with hemispherectomy? And where should we place the intact brain?

If numerosity of mind were a continuum, it might be easier to describe the splitbrain as well as the intact brain. As it is, our current language forces us to jump one way or the other. Between "one mind" and "two minds" to describe either the splitbrain or the intact brain, "two minds" seems closer to being correct.

If the language which we have inherited from the past were changed, we might be able to provide better answers than we have at present. Until we do have a better vocabulary, we are stuck with anthropomorphizing the hemispheres. And we are stuck with the resistance that such usage will elicit in those who (however poorly unified they may be) know that they are better unified than any pair of people with hemispherectomy.

We are on the threshold of a revolution in the way in which we talk about the nature of mind and about human nature. We possess now an abundance of facts which are not adequately comprehended by the concepts which we have inherited from the past. Thinking of the mind as double can provide some resolution of the centuryold conflict in neurology between holism and topism. And it seems to me quite likely that thinking in terms of a duality of the mind can provide us with better understanding of repression and of conscientiation than some of the wellknown theories with which we are presently provided.

We cannot yet say what new concepts (presumably quantitative) shall serve us best. But we do see that they are needed; we can see in part the direction they should take; and we all look forward to the day when the implications of the split-brain research emerge in a form which can help guide human society toward an improved understanding of its own internally conflicted creativity.

REFERENCES

1. Bogen JE. Concluding overview. In: Reeves A, ed. Epilepsy and the corpus callosum. New York: Plenum, 1985.

2. Bogen JE. The other side of the brain 1. Dysgraphia and dyscopia following cerebral commissurotomy. Bull Los Angeles Neurol Soc 1969;34:73105.

3. Bogen JE. The other side of the brain 11. An appositional mind. Bull Los Angeles Neurol Soc 1969;34: 135 162.

4. Bogen JE, Bogen GM. The other side of the brain 111. The corpus callosum and creativity. Bull Los Angeles Neurol Soc 1969;34:191220.

5. Bogen JE. Mental numerosity: Is one head better than two? Behav Brain Sci 1981;4:100101.

6. Wigner ER Symmetries and reflections. Cambridge: M.l.T. Press, 1967:176.

7. Mehra J. Quantum mechanics and the explanation of life. Amer Sci 1973;61:722728.

8. Bohm D. Wholeness and the implicate order. London: Routledge and Kegan Paul, 1981.

9. Sherrington CS. The integrative action of the nervous system. Cambridge: Cambridge Univ. Press, 1947.

10. Chadwick J, Mann WN. The medical works of Hippocrates. Oxford: Blackwell, 1950: 183.

11. Wigan AL. The duality of the mind (originally 1844). J. Simon, Box 4071, Malibu, CA 90265,

12. Jackson JH. On the nature of the duality of the brain. Medical Press and Circular 1: 19,41,63 (1874). (Repninted in Brain 38, 1915 and in J. Taylor (ed.) Selecfed Writings of John Hughlings

13. Harrington A. I9th Century ideas on hemisphere differences and "duality of mind." Behav Brain Sci 1985;8:617-660. 13a.Harrington,A Medicine,Mind and the Double Brain 1987 :Princeton U.Press

14. Sperry RW. Cerebral organization and behavior. Science 1961; 133:1749-1757. .

15. Sperry RW. Lateral specialization in the surgically separated hemispheres. In: Schmitt FO, Worden FG, eds. Neurosciences: Third study program. Cambridge, MA: MIT Press, 1974:519.

16 Sperry RW Some effects of disconnecting the cerebral hemispheres. Science 1982;217:1223-1226.

17. Young RM. Mind, brain and adaptation in the nineteenth century. Oxford: Clarendon Press,1970.

18. Bogen JE. SplitBrain Syndromes. In: Frederiks JAM, ed. Handbook of Clinical Neurology Vol. 45. Amsterdam/New York: Elsevier, 1985.

19. Bogen JE. The callosal syndromes. In: Heilman KM, Valenstein E, eds. Clinical Neuropsychology 2nd Ed. New York: Oxford Univ. Press, 1985.

20. Bogen JE. Partial hemispheric independence with the neocommissures intact. In: Trevarthen C,(ed)Brain circuits and functions of the mind London: Cambridge Univ. Press, 1986.

21. Bogen JE. Hemispherectomy and the placing reactions in cats. In: Kinsbourne M, Smith WL, eds. Hemisphere disconnection and cerebral function. Springfield, IL: Thomas, 1974.

22. Wenzel BM, Tschirgi RD, Taylor JL. Effects of early postnatal hemidecortication on spatial discrimination in cats. Exp Neurol 1962;6:332339.

23. Kruper DC, Patton RA, Koskoff YD. Visual discrimination in hemicerebrectomized monkeys. Physiol Behav 1971;7:173179.

24. Marks CE. Commissurotomy consciousness and unity of mind. Montgomery, VT: Bradford

25. Bogen JE. Further discussion on splitbrains and hemispheric capabilities. Brit J Philosophy Sci

26. Doty RW, Nagrao N, Yamaga K. The unilateral engram. Acta Neurobiol Exp 1973 ;33:711 728.

27. Puccetti R. Brain bisection and personal identity. Brit J Philosophy Sci 1973;24:339-55.

28. Puccetti R. The case for mental duality: Evidence from splitbrain data and other considerations. Behav Brain Sci 1981;4:93123.

29. Trevarthen CB, Sperry RW. Perceptual unity of the ambient visual field in human commissurotomy patients. Brain 1973;96:547-70.

30. Weiskrantz L, Warrington EK, Sanders MD, Marshall J. Visual capacity in the hemianopic field following a restricted occipital ablation. Brain 1974;97:709-28.

31. Gazzaniga MJ, LeDoux JE. The integrated mind. New York: Plenum, 1978.

32. Sperry RW, Zaidel E, Zaidel D. Selfrecognition and social awareness in the deconnected minor hemisphere. Neuropsychologia 1979:17:15366.

33. Magoun HW. The waking brain. Springfield, IL: Thomas. 1958:106.

34. Berlucchi G. 10 Electroencephalographic activity of the isolated hemicerebrum of the cat. Exp Neurol 1966;15:220-28.

35. Berlucchi G. Electroencephalographic studies in "split brain" cats. Electroencephalogr Clin Neurophysiol 1966;20:348-56.

36. Michel F, Roffwarg HR Chronic split brain stem preparation: Effect on the sleepwaking cycle. Experientia 1967;23:12628.

37. Michel F. Sleep and waking in cats with various sagittal sections of the brain. In: Cernacek J, Podivinsky F, eds. Cerebral interhemisphenc relations, Bratislava: Vydavalestvo Slovenskej Akad, 1972.

38. Lilly JC. Comment on cerebral dominance in the dolphin. In: Mountcastle V, ed. Interhemispheric relations and cerebral dominance. Baltimore: Johns Hopkins Univ. Press, 1962.

39. Serafetinides EA, Shurley JT, Brooks RE. Electroencephalogram of pilot whale Globicephala scammoni in wakefulness and sleep: Lateralization aspects. Int J Psychobiol 1972;2:129-135.

40. Mukhametov LM, Supin AY, Polyakova IG. Interhemispheric assymetry of the electroencephalographic sleep pattern in dolphins. Brain Res 1977;134:581-584.

41. Ridgway SH, Flanigan WF. Electrophysiological observation during sleep in the bottlenosed dolphin. 1984;

42. Gazzaniga M. Eye position and visual motor coordination. Neuropsychologia 1969;7:379-382.

43. Myers JJ. Cognitive transfer from nght to left hemisphere after section of the forebrain commissures. Thesis, California Institute of Technology, 1984.

44. Cronin-Golomb A. Intrahemispheric processing and subcortical transfer of nonverbal information in subjects with complete forebrain commissurotomy. Thesis, California Institute of Technology, 1984. .,~ .

45. Butler SR, Norrsell U. Vocalization possibly initiated by the minor hemisphere. Nature 1968;220:793-794.

46. Levy J, Trevarthen C, Sperry RW. Perception of bilateral chimeric figures following hemispheric deconnexion. Brain 1972;95:61-78.

47. ZolaMorgan S, Squire LR, Mishkin M. The neuroanatomy of amnesia: Amygdalahippocampus versus temporal stem. Science 1982;218:1337-1339.

48. Harvey W. Anatomical studies on the motion of the heart and blood. (C. Leake, trans.) Springfield: Thomas, 1958, originally 1628.

49. Garrison FH. History of medicine 3rd Ed. Philadelphia: Saunders, 1922:243.

50. Tomasch J. Size, distribution, and number of fbres in the human corpus callosum. Anat Rec 1954;119:719.

51. Innocenti GM. What is so special about callosal connections? In: Lepore F, Ptito M, Jasper HH, eds. Two Hemispheres, One Brain? New York: Alan Liss, 1986.

52. Berlucchi G. Callosal activity in unrestrained, unanesthetized cats. Arch Ital Biol 1965;103:623-634.

53. Myers RE. Organization of visual pathways. In: Ettlinger EG, ed. Functions of the corpus callosum. Ciba foundation study group No. 20. Boston: Little, Brown, 1965:133143.

54. Berlucchi G. Anatomical and physiological aspects of visual functions of corpus callosum. Brain Res 1972;37:371-392.

55. Doty RW, Negrao N. Forebrain commissures and vision. In: Jung R, ed. Handbook of sensory physiology Vll/3B Berlin: SpringerVerlag, 1973.

56. Newsome WT, Allman JM. Interhemispheric connections of visual cortex in the owl monkey,Aotus trivirgatus, and the bushbaby, Galago senegalensis. J Comp Neurol 1980;194:209-233. ~

57. Marzi A. Transfer of visual information after unilateral input to the brain. Brain Cognition 1986,5:163-173

58. Myers RE. Transmission of visual information within and between the hemispheres: A behavioral study. In: Mountcastle VB, ed. Interhemispheric relations and cerebral dominance. Baltimore, Johns Hopkins Univ. Press, 1962.

59. Semmes J, Mishkin M. A search for the cortical substrate of tactual memories. In: Ettlinger G, ed. Functions of the corpus callosum, London: Ciba, 1965.

60. Butler CR, Francis AC. Splitbrain behavior without splitting. Tactile discriminations in monkeys. Isr. J. Med Sci 1973;9:79-84.

61. Berlucchi G, Buchtel E, Marzi CA, Mascetti GG, Simoni A. Effects of experience on interocular transfer of pattern discriminations in splitchiasm and splitbrain cats. J Comp Physiol Psychol 1978;92:532-543.

62. Sperry RW. Clark E. Interocular transfer of visual discrimination habits in teleost fish. Physiol Zool 1949;22:372-378.

63. Russell IS, Morgan SC. Interocular transfer of visual learning in the rat. In: Russell IS, Van Hof MW, Berlucchi G, eds. Structure and function of cerebral comissures. Baltimore: Univ. Park Press, 1979.

64. Van Hof MW Interocular lransfer and inlerhemispheric communication in the rabbit. ln: Russell lS, Van Hof MW, Berlucchi G, eds. Structure and function of cerebral commissures. Baltimore: Univ. Park Press, 1979.

65. Beritov J, Chichinadze N. On the localization of cortical processes evoked by visual stimulation. Tr Inst Fiziol Beritashvili 1937;3:361-364.

66. Graves JA, Goodale MA. Do training conditions affect interocular transfer in the pigeon? In: Russel IS. van Hof MW, Berlucchi G. eds. Structure and function of cerebral commissures. Baltimore: Univ. Park Press, 1979.

67. Goodale MA, Graves JA. Retinal locus as a factor in interocular transfer in the pigeon. In: Ingle DJ, Goodale MA, Mansfield RM, eds. The analysis of visuai behavior, Cambridge, MA: MIT Press, 1982:211-240.

68. Bell GA. Gibbs ME. Unilateral storage of monocular engram in dayold chick. Brain Res 1977,124 263-270.

69. Benowitz L. Conditions for the bilateral transfer of monocular learning in chicks. Brain Res1974; 65:204-213

70. Greif KF Bilateral memory for monocular onetrial passive avoidance in chicks. Behavioral Biology 1976: 16;453-62

71. Hodge RJ, Gibbs ME, Ng KT. Engram duplication in the dayold chick. Behav Neural Biol 1981 ;31:283 -298.

72. Marzi CA, DiStefano M, Simoni A. Pathways of interocular transfer in Siamese cats. In: Russell IS, Van Hof MWS, Berlucchi G, eds. Structure and function of cerebral commissures. Baltimore: Univ. Park Press, 1979.

73. Butler SR. Interhemispheric transmission of visual information via the corpus callosum and anterior commissure in the monkey. In: Russell IS, Van Hof MW, Berlucchi G, eds. Structure and function of cerebral commissures. Baltimore: Univ. Park Press, 1979.

74. Doty RW. Some thoughts, and some experiments, on memory. In: Butters N, Squire L, eds. The neuropsychology of memory. New York: Guilford Press, 1983.

75. Doty RW, Lewine JD, Ringo JL. Mnemonic interaction between and within cerebral hemispheres in macaques. In: Alkon DL, Woody CD, eds. Neural mechanisms of conditioning. New York: Plenum, 1984.

76. Hamilton CR. Investigations of perceptual and mnemonic lateralization in monkeys. In: Harnad S, Doty RW, Goldstein L, Jaynes J, Krauthamer G, eds. Lateralization in the nervous system. New York: Academic Press, 1977.

77. Hamilton CR. Mechanisms of interocular equivalence. In:lngle DJ, Goodale MA, Mansfield RJW, eds. Analyses of visual behavior. Cambridge: MIT Press, 1982.

78. Zeki SM. Visual deficits related to size of lesion in "prestriate" cortex of optic chiasm sectioned monkeys. Life Sci 1967;6:1627-38.

79. Voneida TJ, Robinson JS. Effect of brain bisection on capacity for cross comparison of patterned visual input. Exp Neurol 1970;26:6071.

80. Abuladze KS. (1963). Functioning of paired organs. Leningrad: Medgiz 1961 (Russian). Translation by Crawford, R. New York: Macmillan (cited by Doty and Negrao, 1973).

81. Kaas J, Axelrod S, Diamond IT. An ablation study of the auditory cortex in the cat using binaural tonal patterns. J Neurophysiol 1967;30:710724.

82. Bures J, Buresova O, Krivanek J. The mechanism and application of Leão's spreading depression of electroencephalographic activity. Prague: Academia, 1974.

83. Denenberg VH. Micro and macro theories of the brain. Behav Brain Sci 1983;6:174178.

84. Luria AR. Traumatic aphasia. Acad. Med. Sci., Moscow; English edition: Mouton, the Hague.

85. Marie, R and Foix, C. Les aphasies de guerre. Rev Neurol i917;31/32:53-87.

86. Russell WR, Espir MLE. Traumatic aphasia. London: Oxford Univ. Press, 1961.

87. Newcombe F. Missile wounds of the brain: A study of psychological deficits. London: Oxford Univ. Press, 1969.

88. Milner B. Interhemispheric differences in the localization of psychological processes in man. Brit Med Bull 1971;27:272-277.

89. Milner B. Clues to the cerebral organization of memory. In: Buser PA, RougeulBuser A, eds. Cerebral correlates of conscious experience. Amsterdam: NorthHolland, 1978.

90. Smith A. Dominant and nondominant hemispherectomy. In: Kinsbourne M, Smith WL, eds. Hemispheric disconnection and cerebral function. Springfield. IL: Thomas, 1974:5-33.

91. Petrinovich L. Cortical spreading depression and memory transfer: A methodological critique. Behav Biol 1976;16:79-84.

92. Gaston KE. Brain mechanisms of conditioned taste aversion learning: A review of the literature. Physiol Psychol 1978:6:340353.

93. Yakovlev Pl. Lecours AR. The myelogenetic cycies of regional maturation of the brain. In: Minkowski A. ed. Regional development of the brain in early life. Edinburgh: Blackwell. 1967.

94. Galin D. Johnstone J. Nakell L. Herron J. Development of the capacity for tactile information transfer between hemispheres in normal children. Science 1979;204:1330-1332.

95. Risse GL. Gazzaniga MS. Wellkept secrets of the right hemisphere: A carotid amytal study of restricted memory transfer. Neurology 1978:28:950-953.

96. Lesser RP, Dinner DS. Luders H. Morris HH. Memory for objects presented soon after intracarotid amobarbital sodium injections in patients with medically intractable complex partial seizures. Neurology i986:36:895-899.

97. Harris LJ. Which hand is the ''eye'' of the blind?óA new look at an old question. In: Herron J,(ed.) The neuropsychology of lefthandedness. New York: Academic Press. 1980. ~ ~i"~

98. Hicks RE, Frank JM, Kinsbourne M. The locus of bimanual skill transfer. J Gen Psychol. 1982;107 277-281

99. Goodglass H, Calderon M. Parallel processing of verbal and musical stimuli in right and left hemispheres. Neuropsychologia 1977;15:397-407.

100. Sidtis JJ. Bryden MP. Asymmetrical perception of ]anguage and music: Evidence for independent processing strategies. Neuropsychologia 1978;16:627-632.

101. Bradshaw JL. Nettleton NC. Human cerebral asymmetry. Englewood Cliffs, NJ: PrenticeHall, 1983.

102. Landis T. Assal G. Perret E. Opposite cerebral hemispheric superiorities for visual associative processing of emotional facial.expressions and objects. Nature 1979;278:739-740.

103. Campbell R, Landis T. Regard M. Facial recognition and lipreading, a neurological dissociation. Brain 1986;109:509-521.

104. Landis T, Graves R, Goodglass H. Dissociated awareness of manual performance on two different visual associative tasks: A "splitbrain" phenomenom in normal subjects? Cortex 1981 ;17:435-440.

105. Desmedt JE. Mise en evidence d'electrogeneses specifiques de l'hemisphere droit nondominant chez l'homme intact. Bull L'Academie Nationale Medecine 1977;161:623-626.

106. Desmedt JE. Active touch exploration of extrapersonal space elicits specific electrogenesis in the right cerebral hemisphere of intact righthanded man. Proc Natl Acad Sci USA 1977; 74:4037- 4040.

107. Zaidel E. Disconnection syndrome as a model for laterality effects in the normal brain. In: Hellige JB, ed. Cerebral hemisphere asymmetry: Method, theory and application. New York: Praeger, 1983:95-151.

108. Berlucchi G. Two hemispheres but one brain. Behav Brain Sci 1983;6:171-172.

109. Bryden MP Laterality: Functional asymmetry in the intact brain. New York: Academic Press. 1982.

110. Corballis MC. Human laterality. New York: Academic Press, 1983. .

111. Trevarthen CB Hemispheric specialization. Handbook of physiology-The nervous system lll Chap. 25 Washington, DC Amer Physiol. Soc 1984

112. Gordon H. The neurobiological basis of hemisphericity. In: Trevarthen C, ed. Brain circuits and theories of mind.Cambridge:CambridgeUniv.Press, 1985.

113. Holton G. Thematic origins of scientific thought. Cambridge: Harvard Univ. Press. 1973. '

114. Bohr N. Atomic physics and human knowledge. New York: Wiley, 1958.

115. Walter DO. Alternatives to continuity, observability, and passivity in biological modeling: tribute to McCulloch. Math Biosci 1971:8594.

116. Walter DO. Obiectivity revised: Models from 20thcentury physical science for use in the human sciences. Int J Neurosci 1971:1:243-249.

117. Feynman RP. Leighton RB. Sands M. The Feynman lectures on physics. Reading. MA: AddisonWesley. 1963.

118. Heisenberg W. The physical principles of the quantum theory. New York: Dover. 1949.

119. Rohrlich F. Facing quantum mechanical reality. Science 1983: 221:1251-1255.

12(). Putterman S. Impossibility of observing coherent macroscopic quantum superposition Phys Lett 1983:98A:324-328.

121. Iberall AS. On complementary-a rebuttal. J Soc Biol Structures 1979:2:173-174.

122. Kadanoff LP Roads to chaos. Phys. Today. 1983:46-53.

BULLETIN OF CLINICAL NEUROSCIENCES 51, 3 - 29 (1986)