In conformity with a rule which seems applicable to every science of observation… it is the exceptional phenomenon which is likely to explain the usual one; and, consequently, whatever we can observe that has to do with invention, is capable of throwing light on psychology in general. (Hadamard, 1949, p. 136)
I think that a materialist definition of genius is impossible, which is why the idea of genius is so discredited in an age like our own, where materialist ideologies predominate. (Bloom, 2002, p. 12)
In his beautiful memorial tribute to F. W. H. Myers, William James (1901) endorsed Myers’s broad vision of a scientific psychology, contrasting it with the much narrower vision produced by a different type of scientific imagination, the “classic-academic” type, which he describes in terms almost equally applicable to present-day mainstream cognitive psychology:
The human mind, as it is figured in this literature, was largely an abstraction. Its normal adult traits were recognized. A sort of sunlit terrace was exhibited on which it took its exercise. But where that terrace stopped, the mind stopped; and there was nothing farther left to tell of in this kind of philosophy but the brain and the other physical facts of nature, (p. 14)
Thanks to Myers and other “romantic improvers,” however, “a mass of mental phenomena are now seen in the shrubbery beyond the parapet,” and “the world of mind is shown as something infinitely more complex than was suspected” (p. 14). James explicitly characterizes Myers as the leader of this expansionist movement: “Through him for the first time, psychologists are in possession of their full material, and mental phenomena are set down in an adequate inventory” (p.16). Moreover, Myers was more than just a prodigious collector; he had shown a Darwin-like genius for organizing and coordinating this mass of material in service of what James regarded as the first serious scientific attempt to delineate the constitution of this transmarginal or subliminal background of the mind. This problem—henceforward to be known as “the problem of Myers”—”still awaits us as the problem of far the deepest moment for our actual psychology, whether his own tentative solutions of certain parts of it be correct or not” (p. 18).
Chapter 2 presented a general account of Myers’s theory, but here we will concentrate in much greater detail on just one of the empirical topics that Myers himself dealt with explicitly—genius. This topic, we will argue, provides another important arena within which Myers’s views, when revisited in light of the much larger body of relevant evidence available today, can be recognized as preferable in a variety of ways to the alternatives that mainstream psychological theory and research have so far provided.
Genius is obviously a topic of great human interest and significance, to all of us as individuals and to civilization as a whole. It also seems self-evident, as Wind’s principle and our first epigraph suggest, that a cognitive science capable of accommodating the fully-developed phenomena of genius would be capable a fortiori of accommodating cognitive phenomena in general. However, despite this seemingly crucial importance of genius as a kind of benchmark and navigational aid for progress in scientific psychology, its treatment to date for the most part reflects the general history of 20th-century psychology, as summarized in our Introduction, and has been anything but satisfactory or illuminating.
In the first four decades following the rise of Watson’s radical behaviorism, for example, the mainstream American psychological literature was practically devoid of relevant studies. In his landmark Presidential Address to the American Psychological Association J. P. Guilford (1950) reported that only 186 of the roughly 121,000 entries in Psychological Abstracts up to that date had addressed this vital topic. Guilford bemoaned this “appalling” neglect, and his challenge provoked a modest increase in research output which has continued and even slowly intensified up to the present. Nevertheless, even now such research represents a relatively tiny and specialized sub-field of psychology. Moreover, by far the greatest proportion of research to date has been carried out in the framework of Guilford’s own psychometric tradition, dating back to Galton and Binet, which revolves mainly around measurement of “divergent thinking,” “fluency,” “flexibility,” puzzle-solving behaviors, and the like, typically in student volunteers.
We do not wish to disparage unduly the modern research tradition, and we hasten to add that it includes many other threads, such as the early work of Gestalt psychologists on “insight,” psychoanalytic investigations of “primary process” thinking, and detailed studies of life history and personality characteristics in demonstrably productive individuals, some of which we will touch upon later.
For purposes of this introduction, however, the important bottom-line fact is that even to many of its practitioners, not to mention outside observers such as ourselves, modern “creativity” research appears mired, overall, in a rather dismal state. A number of the contributors to the recent Handbook of Creativity (R. J. Sternberg, 1999)—the explicit goal of which was “to provide the most comprehensive, definitive, and authoritative single volume review available in the field”—candidly acknowledge this sorry state of affairs, which also is fully apparent in the most current state-of-the-art review consulted by us (Runco, 2004). In effect, we suggest, the study of the real thing—”genius”—has largely degenerated in modern times into the study of diluted cognates such as “creativity” or even “talent,” which happen to be relatively accessible to the more “objective” means of investigation currently favored by most investigators.
Myers consciously and deliberately took an approach targeted throughout to genius in its fullest expressions—what “the highest minds have bequeathed to us as the heritage of their highest hours” (HP, vol. 1, p. 120). We will show that by means of this approach, honoring Wind’s principle, Myers produced an account of genius which anticipates to a remarkable degree most of what has been best in more recent work, while also accommodating various unusual phenomena, such as psychological automatisms and secondary streams of consciousness, altered states of consciousness, unusual forms of symbolic thinking, and psi, that are inescapably bound up with this topic but scarcely touched upon in contemporary mainstream accounts. His views also have been confirmed at various important points by more recent empirical and theoretical investigations, and they have important implications for the further evolution of scientific psychology more generally.
To substantiate these claims we will present Myers’s account of genius in considerable detail, situating it as we proceed with respect to the main relevant trends in more recent psychological research. Our presentation is necessarily telegraphic and selective, for the existing literature on “creativity,” though but a tiny proportion of the psychology literature as a whole, harbors something upwards of 10,000 papers and books; our purpose is not to survey this enormous literature comprehensively, but to advance and defend a particular point of view.
Myers’s chapter on genius in Human Personality is almost identical to his earlier paper on this subject (Myers, 1892d). The chapter itself is conspicuously atypical in form, in that it is very short and less densely empirical than its siblings, and lacks an appendix. Yet there can be no doubt that it both illustrates and provides support for some of Myers’s most central and deeply-felt convictions regarding the subliminal realm and its role in human mental life. His chapter on genius is itself a work of genius, so full of bold and challenging observations and speculations that it will require the balance of our present chapter and all of the next to do it reasonably full justice.
Myers begins by setting forth the general character of his conception of genius. In his previous chapter he had characterized hysteria as a disintegrative or “dissolutive” process involving loss of control of normally supraliminal elements of the personality. Genius for Myers presents the opposite situation. Specifically, in genius an increased “strength and concentration of the inward unifying control” (HP, vol. 1, p. 70) results in enhanced coordination and integration of the supraliminal and subliminal phases of personality. In effect genius stands in relation to ordinary personality roughly as ordinary personality itself stands in relation to hysteria. Genius represents the evolution of personality toward a more ideal form of psychic functioning, and therefore toward a truer standard of “normality.”
In taking this position Myers directly opposed a sizeable cadre of writers including Max Nordau, John Nisbet, and Cesare Lombroso, who at that time were busily engaged in deflating “genius” by characterizing it as nothing more than a form or manifestation of “degeneracy” or “madness” of one or another supposedly well-understood type.1
Myers’s basic response to this deflationary movement, then in full flood, incisively encapsulates the main features of his own very different view (HP, vol. 1, p. 71):
On this point I shall join issue; and I shall suggest, on the other hand, that Genius—if that vaguely used word is to receive anything like a psychological definition—should rather be regarded as a power of utilising a wider range than other men can utilise of faculties in some degree innate in all;—a power of appropriating the results of subliminal mentation to subserve the supraliminal stream of thought;—so that an “inspiration of Genius” will be in truth a subliminal uprush, an emergence into the current of ideas which the man is consciously manipulating of other ideas which he has not consciously originated, but which have shaped themselves beyond his will, in profounder regions of his being. I shall urge that there is here no real departure from normality; no abnormality, at least in the sense of degeneration; but rather a fulfilment of the true norm of man, with suggestions, it may be, of something supernormal’,—of something which transcends existing normality as an advanced stage of evolutionary progress transcends an earlier stage.2
Note first Myers’s insistence on a psychological definition of genius. Some later investigators have focused instead primarily on its objective manifestations, so that “genius” becomes an inference backwards from the appearance of products that are regarded as useful or valuable or satisfying by some audience at some point in time. In the work of Lange-Eichbaum (1932), representing this tendency of thought at its extreme, the personality of the genius essentially vanishes, the term “genius” implying only a relation of approval between products of creative activity and the relevant community of evaluators. The recently articulated “systems perspective” of Csikszentmihalyi (1988, 1996) and others similarly downplays the psychological dimensions of the subject in favor of cultural, social, and situational factors.
Myers, however, deliberately takes the opposite tack. He first clearly acknowledges that a work of genius conventionally satisfies two distinct sorts of requirements: “It must involve something original, spontaneous, unteachable, unexpected, and it must also win for itself in some way the admiration of mankind” (HP, vol. 1, p. 75). But he immediately makes clear that his concern is with the first sort of criterion, which corresponds to a real psychological class. As he puts it:
What the poet feels while he writes his poem is a psychological fact in his history; what his friends feel while they read it may be a psychological fact in their history, but does not alter the poet’s creative effort, which was what it was, whether any one but himself ever reads his poem or no.
Thus for example Hartley Coleridge seemed to Myers to have manifested the characteristic psychological marks of genius even though he never produced anything tangible that commanded wide attention.
Myers’s central concern and goal, in sum, is to elucidate the psychological processes involved in undeniable manifestations of genius, and to discern their broader implications for the theory of human personality. We will present and discuss his more detailed arguments under these headings.
A traditional descriptive model of the creative process, based on the self-observation and testimony of large numbers of variously eminent persons, provides a useful organizing framework for this discussion. Credit for explicitly formulating this model is usually given to Graham Wallas (1926), a political scientist and administrator primarily concerned with pedagogical matters, but it was also formulated in nearly identical terms and in greater detail by psychologist Eliot Dole Hutchinson (1931, 1939). The model posits four phases or stages that can often be discerned in high-level creative effort: (1) preparation; (2) incubation; (3) illumination; and (4) verification. Briefly, preparation refers primarily to the initial stages of intense voluntary effort on a particular work or problem (although it is sometimes generalized to include the typically lengthy period of time in which high-level technical skills relevant to the task are laboriously acquired). If this initial effort fails, the work or problem may temporarily be put aside in frustration, this being the stage of incubation or renunciation, in which conscious effort seems to be largely or wholly absent. Something more than simple rest or dissipation of inhibitions seems to be involved during the incubation period, for then comes illumination, inspiration, or insight, in which radically new ideas intrude into consciousness, often suddenly, copiously, and with strong accompanying affect. This leads to a further stage of voluntary effort, verification, in which the new material may be evaluated, elaborated, and worked into the structure of the evolving product.3
Only in rare cases can these “stages” accurately be pictured as unrolling in a rigid sequence, or at a single level. Normally they go forward concurrently at multiple, overlapping, and recursively interacting levels in regard to different elements of the total problem situation. The incubation and inspiration stages, which form the psychological heart of the model, are not invariably or necessarily discernible in individual episodes of creative activity, and even when recognizably present they can vary widely in duration, character, and importance (Ghiselin, 1952; Hutchinson, 1931, 1939; Koes-tler, 1964; Wallas, 1926).
A descriptive model of this general sort is accepted by nearly everyone who has studied the subject of genius seriously. There is, however, a modern “deflationary” tradition, exemplified by Weisberg (1986, 1999) and Perkins (1981, 2000), that we must comment upon at least briefly before proceeding. This school wishes to deny that there is anything special or unusual about the character of the mental processes occurring in genius-level creativity, and in particular to deny that genius ever depends upon or utilizes any special sort of “unconscious” work. Nothing more is involved, on this view, than unusually tenacious, disciplined, and incremental application of cognitive processes of the ordinary sort. Not surprisingly, this contrarian view arises mainly in the context of laboratory studies of puzzle-solving behaviors and other sorts of on-demand “creativity,” mostly in student volunteers. The fact that the alleged phenomena of incubation and illumination cannot readily and reliably be evoked under their preferred laboratory conditions disposes such workers to conclude that the phenomena do not exist, or are of only minor importance. Their existence and importance in any event cannot be established by first-person testimony from any number of witnesses, no matter how eminent, because such testimony is merely “anecdotal” and always subject to corruption by well-known hazards such as limitations of observation, failures of memory, motivated distortions in service of public personae, and so on.
With all due respect, this “nothing-special” view of genius seems to us a particularly egregious example of “methodolatry” (see Introduction), coupled with willful disregard of an enormous and fundamentally coherent mass of evidence deriving from persons especially well positioned and qualified, generally speaking, to observe and report the phenomena of greatest psychological interest. As Ghiselin (1952) remarks, “the fundamentals… are all but inescapable” (p. 11). The unwillingness of many modern literary scholars to confront genius in its full-blown phenomenological reality has recently provoked Harold Bloom (2002) to castigate them as “cultural levelers, quite immune from awe” (p. 7), and in our view that judgment applies equally here. We also find it especially ironic, and revealing, that in the end both Weisberg and Perkins are forced to acknowledge that even the “ordinary” cognitive operations to which they aspire to reduce all manifestations of genius—operations such as remembering, recognizing, realizing, seeing-as, noticing, and reasoning—remain themselves at bottom (and again in accordance with Wind’s principle) unexplained. It is also ironic, of course, that their repudiation of the possibility of unconscious mental work in genius comes just as many other mainstream psychologists are re-admitting such things in the form of the “cognitive unconscious.” As in Chapter 5, the real issue before us therefore becomes that of the scope and character, rather than the existence per se, of the mental activity that goes on behind the scenes, outside the purview of ordinary waking consciousness.
Let us presume, then, that the Wallas/Hutchinson model is broadly correct, descriptively. Myers directs his analysis straight to the psychological core of this model, to its most critical, mysterious, and controversial component—specifically, to the phase of “illumination,” or in his own more theory-laden terminology, “subliminal uprush.” Three main features of his analysis—continuity, automatism, and incommensurability—will command the bulk of our attention.
Myers emphasizes the fundamental continuity between the mental processes at work in genius and those of more everyday character. He is a thoroughgoing naturalist, and for him genius is first and foremost an intensification of phenomena already observable in germ in the central, supraliminal part of the mental “spectrum” (Chapter 2), rather than some sort of supernatural gift of faculty altogether new. Some degree of cooperation between supraliminal and subliminal elements of personality, he points out,
is constantly occurring on a smaller scale in the inner life of most of us. We identify ourselves for the most part with a stream of voluntary, fully conscious ideas,—cerebral movements connected and purposive as the movements of the hand which records them. Meantime we are aware also of a substratum of fragmentary, automatic, liminal ideas, of which we take small account. These are bubbles that break on the surface; but every now and then there is a stir among them. There is a rush upwards as of a subaqueous spring; an inspiration flashes into the mind for which our conscious effort has not prepared us. This so-called inspiration may in itself be trivial or worthless; but it is the initial stage of a phenomenon to which, when certain rare attributes are also present, the name of genius will naturally be given. (HP, vol. 1, p. 77)
As described in Chapter 2, Myers conceives of the subliminal as being in all of us restrained or released in accord with dynamic adjustments in the level of some sort of “threshold”—or in the permeability of a membrane or “diaphragm”—which regulates its supraliminal expression. In most of us this dynamic process is evidenced by occurrences such as nocturnal dreams and the mild dissociative fluctuations, such as daydreams and hypnagogic states, that mark daily life. In genius, however, these dynamic adjustments are somehow amplified, providing correspondingly greater supraliminal access to products or elements of subliminal mentation.
Not all such products are of equal value, however, for “hidden in the deep of our being is a rubbish-heap as well as a treasure-house” (HP, vol. 1, p. 72). Myers takes pains to guard against any naive or unchecked romanticism:
I shall be obliged in this chapter to dwell on valuable aid rendered by subliminal mentation; but I do not mean to imply that such mentation is ipso facto superior to supraliminal, or even that it covers a large proportion of practically useful human achievement. When I say “The differentia of genius lies in an increased control over subliminal mentation,” I express, I think, a well-evidenced thesis, and I suggest an important inference, namely, that the man of genius is for us the best type of the normal man, in so far as he effects a successful co-operation of an unusually large number of elements of his personality—reaching a stage of integration slightly in advance of our own. This much I wish to say: but my thesis is not to be pushed further:—as though I claimed that all our best thought was subliminal, or that all that was subliminal was potentially “inspiration.” (HP, vol. 1, p. 72)
For Myers the subliminal realm has a complex hierarchical organization marked by successively higher levels of integration, each with its own characteristic functional properties or “adits and operations.” In deliberate analogy with Hughlings Jackson’s evolutionary model of hierarchical organization in the human nervous system (see our Chapter 2), Myers pictures the subliminal tentatively and broadly as comprising three such levels: The lowest involves the “minimal psychic concomitant, whatever that may be” (HP, vol. 1, p. 74), of bare vegetative functions of the sort governed by Jackson’s lowest-level physiological centers, as well as mechanical effects due to habit, adaptation, stimulus inputs that escape conscious detection, and the like. A middle region, the “hypnotic stratum,” is associated with automatisms such as the dissociative phenomena of hysteria and deep hypnosis, but it also supplies the content of ordinary dreaming, daydreaming, and imagining and is the source of the “mythopœic” function, an incessant “strange manufacture of inward romances” (HP, vol. 2, p. 130).4 The deepest region is the least known and hardest to describe, but is above all the locus of various forms of supernormal contact with the outside world, including both psi phenomena and the intuitions and inspirations of genius.
These “strata” are not to be conceived in quasi-geological fashion as if they are static, immobile, or rigidly separated by impassable barriers: “They are strata (so to say) not of immovable rock, but of imperfectly miscible fluids of various densities, and subject to currents and ebullitions which often bring to the surface a stream or bubble from a stratum far below” (Myers, 1892b, p. 307). Thus, although the metaphor of “depth” used by Myers and other theoreticians of the subliminal almost inevitably suggests an essentially “geographical” interpretation, for Myers its principal connotations are meant to be functional rather than spatial. Within this basic functional scheme, engagement of the “deepest” strata of subliminal activity is specifically associated for Myers with those profoundest inspirations of genius in which the Self or Individuality comes closest to full and harmonious manifestation of its capabilities, including its supernormal capabilities (HP, vol. 1, p. 73).
Inspirations of genius in general involve successful appeal to the deeper subliminal levels, with success being due to some combination of intensity in the appeal itself (via the preparatory labor) and a favorable psychological constitution (traits) or conditions (states) that provide an unusual “permeability” or openness to the subliminal. Myers’s views here are very similar to those of later observers such as Wallas (1926), Hutchinson (1931, 1939), Ghiselin (1952), and Kubie (1958), among many others. Hutchinson (1939), for example, comments that the characteristic phenomenology of illumination is most pronounced in connection with cases of especially brilliant insight, “where the degree of difficulty and frustration is great and the drive toward accomplishment persistently strong” (p. 232). Myers also emphasizes that the phenomenology intensifies in proportion to “depth.” He focuses in particular on two further psychological characteristics of creative inspiration, both of which flow from, and support, his broader theoretical picture.
Myers holds that the subliminal uprushes of genius belong to the more general category of psychological automatisms (see our Chapter 5) and therefore are inevitably associated not only with automatism itself but with related phenomena such as mediumistic trance and kindred altered states of consciousness. He develops this picture, characteristically, in stages.
Myers turns first to a discussion of “calculating boys,” who for him illustrate the essential psychological workings of higher forms of genius in a usefully “diagrammatic,” verifiable, and semi-quantitative form.5 The scant information then available in regard to some 15 such cases either previously published or known to him personally was sufficient to reveal the affinity of this “computative gift” to other phenomena of subliminal origin, such as hallucinations, rather than to products of ordinary voluntary effort. For example, despite its seemingly necessary connections with more general mathematical knowledge and insight, it is distributed almost at random, appearing among persons of ordinary or even extremely low intelligence as well as in mathematical geniuses such as Gauss and Ampere. It also has a “critical-period” aspect, tending to appear and disappear suddenly in childhood, and if it disappears without being integrated into the general pattern of voluntary skills, it usually leaves behind no memory whatsoever of the processes involved. Most significantly, perhaps, among the dull prodigies it usually operates in the apparent absence of steady conscious effort, or even while the calculator is consciously occupied with other matters. The answer simply appears, usually though not always visually, all at once and with no trace of the steps or processes that led to it. Indeed, the calculator may not even be able to grasp the elementary arithmetical operations that an ordinary person would use to solve, much more laboriously, the same problem. Myers also noted that there was a possible hint of special involvement of the right hemisphere in these phenomena, inasmuch as the two cases for which relevant information was available showed a pronounced tendency toward ambidexterity.
Myers’s observations on prodigious but retarded calculators such as Dase and Fuller have been sustained and generalized in the subsequent century, primarily through further study of what is now called the “savant syndrome,” in which islands of considerable or sometimes spectacular ability appear in the midst of otherwise generalized and profound disability. A particularly valuable survey is that of Treffert (1989), who estimates that something like 100 such “prodigious” cases of the sort that interested Myers have so far been reported, of whom roughly one or two dozen, including “the twins” of Sacks (1987), are currently living. In addition to prodigious calculators, there have been prodigious mechanical, artistic, and especially musical savants, all of whom characteristically display narrow but deep attention coupled with extraordinary memory.6 Treffert speculatively links the savant syndrome to hypothetical abnormalities in development and functional organization of memory and attentional systems in the savant brain, and sensibly calls for anatomical and functional imaging studies (which to our knowledge are just getting underway in earnest) to investigate these. But in the end, echoing Penfield (1975), he acknowledges misgivings as to whether the brain alone can provide the full answer. Both Treffert and Sacks are skilled and caring clinicians with long and first-hand experience of the savant syndrome, and both remain—appropriately, in our opinion—openly awed by it. They also share with Myers a recognition of its deep connections with the mystery of genius.
Myers next presses on to consider more briefly some further instances of apparent subliminal cooperation, instances which share with arithmetic calculation the properties of definiteness and verifiability of result, but which involve sensory phenomena of various sorts. His first examples concern “perceptions of a less specialized kind which underlie our more elaborate modes of cognizing the world around us”—specifically the sense of the passage of time, and the sense of weight or muscular resistance, which rank “among the profoundest elements in our organic being” (HP, vol. 1, p. 85). He cites several cases, drawn mainly from “a sane and waking person” whose acquaintance he had made, in which correct knowledge of some objective state of affairs such as the time of day, or the weight of an animal hide, had appeared spontaneously and in quasi-hallucinatory form, in advance of verification and accompanied by a strong sense of conviction. Myers suggests that such occurrences are best explained as the result of some sort of subliminal calculation, analogous to those of arithmetical prodigies, rather than to any sort of direct or supernormal knowledge.
Hypnotism provides many further analogies to the process of subliminal uprush in genius, as Myers (1898b) had himself argued earlier in an address to the British Medical Association. When a deeply hypnotized girl produces the suggested hallucination of a black cat, for example, she reveals a degree of creativity comparable to that of her dreams but normally unavailable to her waking consciousness. Thus, comments Myers, “as the Sistine Madonna was to Raphael, so to the hypnotized girl is the delusive cat.” Additional examples are provided by successful posthypnotic suggestions, in which the suggested thoughts or actions emerge into consciousness or behavior, suddenly and involuntarily, upon the appearance of a prespecified but consciously forgotten cue. Of special interest among these in the present connection are cases in which the “cue” consists in the passage of a specified amount of time. Careful work by Myers’s colleagues Gurney, Delboeuf, and especially Bramwell had shown that in deeply hypnotizable subjects this sense of the passage of time can sometimes reach astonishing levels of precision, with errors on the order of one part in 2000 or better over intervals of many thousands of minutes.7 McDougall (1911/1961, pp. 353–354) comments favorably on this work, and mentions in a footnote that he had himself encountered the same phenomenon—interestingly enough, in a highly hypnotizable subject who had also proven capable of producing blisters and extravasation of blood from the skin. McDougall further points out that this phenomenon poses a severe challenge to any attempted explanation based on unconscious monitoring of physiological rhythms, because “we know of no bodily rhythm sufficiently constant to serve as the basis of so accurate an appreciation of duration as would have enabled the subject to carry out the suggestion with the high degree of accuracy shown” (p. 353). To our limited knowledge of research on biorhythms, this argument remains valid today, but little further work, unfortunately, appears to have been carried out along these lines.8
Myers turns next to Sir John Herschel, the former royal astronomer, for an example of subliminal products of visual type. Herschel, it turns out, was subject to vivid and kaleidoscopic visual imagery of a highly regular and geometric sort. This imagery occurred involuntarily, usually when he was lying awake in darkness (though twice in broad daylight and twice also in conjunction with chloroform anesthesia for minor surgeries), and was not accompanied by illness or discomfort. Herschel himself had interpreted these hallucinatory phenomena as providing “evidence of a thought, an intelligence, working within our own organization distinct from that of our own personality” (HP, vol. 1, p. 88). Myers agreed, viewing them as expressions of an “indwelling general perceptive power” which we all have, and credited Herschel with having thus originated at least in germ his own more comprehensive theory of subliminal consciousness.
Contemporary research on hallucinatory syndromes lends support to Myers’s conception, although the situation now looks more complex and interesting. To begin, Herschel himself can now be recognized as a migraineur, one of the roughly 1% of the total population who experience migraine aura without accompanying headache (Sacks, 1999, chap. 17; Wilkinson, 2004). Second, “geometrical spectres” like those of his migraine aura are now known to occur under a wide variety of additional circumstances including in particular “hypnagogic” or “twilight” states at sleep onset and offset, the intoxications produced by various hallucinogenic substances such as mescaline and LSD, psychotic breakdowns, fever-induced delirium, stroboscopic visual stimulation at critical frequencies, and sensory deprivation (Klüver, 1966; Sacks, 1999; Siegel, 1977).
The most primitive or elementary forms of such activity typically involve single or multiple points or spots of light, possibly colored, that may organize progressively into a variety of characteristic and simple geometries. Drawing mainly upon his own studies with mescaline, Klüver (1966, p. 66) identified the principal building blocks, or “hallucinatory form constants,” from which more complicated imagery might subsequently evolve. In his original analysis these were just four in number: (1) gratings, lattices, fretworks, filigrees, honeycombs, or chessboards; (2) cobwebs; (3) tunnels, funnels, alleys, cones, or vessels; and (4) spirals.9
The central impulse of mainstream reductionist science, exemplified with particular clarity in the cases of migraine aura (Sacks, 1999; Wilkinson, 2004) and hypnagogic imagery (Mavromatis, 1987; D. Schacter, 1976), has always been to “explain” these recurrent formal characteristics of visual hallucinatory experience as direct expressions in consciousness of events, structures, or processes within the visual pathway. Early accounts of hypnagogia, for example, were relentlessly “peripheralist,” and sought to trace most if not all of the reported phenomenology to “entoptic” sources—that is, to events or structures within the eye itself. Although such sources certainly do play some role in the formation of hypnagogic imagery, presumably by triggering or supplying raw materials for some more central constructive process, few if any now take seriously the notion that hypnagogic imagery, even in its most elementary forms, is in effect generated within the eye.
This general discussion has recently reached a more interesting stage in the context of the migraine aura. The core phenomenology of the aura consists of parallel groups of extremely bright, jagged, scintillating lines, sometimes bounding closed regions described as “fortifications” that expand and move with increasing speed toward the edge of the visual field in a highly stereotypical manner, usually with a region of “scotoma” or temporary blindness trailing behind.10 Early thinking emphasized “static” properties of the aura, viewing it essentially as a kind of phenomenological photograph of the microanatomy of visual cortex, but Lashley (1941), himself a migraineur, introduced the currently dominant “dynamic” view: Specifically, careful comparisons between the evolution of his own aura phenomenology and the known geometry and organization of primary visual cortex led him to hypothesize that the aura results from a traveling wave of neural excitation/depression moving through the cortex at a constant rate of about 3 mm/minute. More generally, the modern idea is that the dynamics of migraine aura phenomenology are to be explained entirely in terms of time-dependent interactions between a spreading-depression process and the functional architecture of visual cortex. Thus for example the phenomenology of jagged boundaries, with short line segments joining at sharp angles, is thought to reflect the classic Hubel/Wiesel architecture of feature detectors, orientation columns, and the like in primary visual cortex (Sacks, 1999; Wilkinson, 2004).
In recent years Lashley’s converging-operations approach has been enriched by the advent of functional neuroimaging and computer modeling methods, both of which have provided additional support for a spreading-wave model (Bressloff, Cowan, Golubitsky, Thomas, & Wiener, 2001; Dahlem, Engelmann, Lowel, & Muller, 2000; Hadjikhani et al., 2001). That some sort of abnormal pattern of physiological activation is present in visual cortex during migraine aura is now widely accepted, although details of the abnormality and its pathophysiological origins remain to be elucidated.
Much less certain, however, is the precise role of this abnormal activity pattern in production of the conscious experience of aura. There are undoubtedly some who, like the early “entoptic” theorists of hypnagogia, imagine visual experience in general to be directly generated by or isomorphic with activity in the visual pathway itself, but we now know that the reality is more complicated. Under normal circumstances, activity in visual cortical areas is in general necessary but not sufficient for the corresponding elements of conscious visual experience (e.g., Dehaene & Naccache, 2001; Haynes & Rees, 2005; Kamitani & Tong, 2005). Something further upstream must interpret or take account of the input-driven activity pattern, which thus initiates, constrains, and guides, but does not fully determine, the conscious experience.
Several further considerations point in the same direction. It may be doubted, for example—and indeed Sacks (1999, p. 289) himself does doubt—whether the standard traveling-wave model really can account in detail for the full-blown phenomenology of the aura. The strain becomes particularly evident as soon as one goes beyond the abstract notion of form constants or other carefully selected aura properties such as the size and speed of motion of “fortifications.” Klüver had noted that his form constants are not restricted to the visual modality, and must therefore have a more central origin. Thus for example one can experience a cobweb on one’s tongue, or feel one’s legs turning into a spiral. Klüver pointed out a number of further properties of hallucinatory experience, moreover, and he himself acknowledged that these more complex phenomena are invariably of far greater salience and impressiveness to the subjects of such experience.
The elementary structures identified as visual form constants, for example, are typically elaborated into ever more complex patterns by an endless and dynamic geometrization, with the elementary forms repeating themselves indefinitely at multiple spatial scales. The boundaries of forms are often other forms. There may be microscopic clarity of detail, with lines almost vanishingly thin, and a kaleidoscopic play of color, brightness, and movement in the evolving structures. Colors themselves may be unnaturally bright, rich, and textured. These visions sometimes can be viewed with the eyes open, in which case they may superimpose upon, alter, or even replace the visual scene itself, in part or in whole, and they often change dramatically in response to tactile or auditory stimuli. Particularly in cases other than migraine aura, the hallucinatory process may generate full-blown dreamlike scenes or stories, sometimes involving additional sensory modalities and incorporating locations, events, or persons long forgotten (Siegel, 1977). As in dreams themselves, this imaginative activity is truly creative and goes far beyond simple literal reproduction of past experience. For example one may observe oneself carrying out novel activities of various sorts, or observe familiar scenes from unfamiliar perspectives.11
In short, in migraine aura as well as the other circumstances in which hallucinatory activity appears, some process is triggered or released which under varying degrees of voluntary control and in service of its ongoing construction of phenomenological reality can utilize, reinterpret, or even ignore (as in negative hallucinations) patterns of neural activity appearing in visual cortex.
Further support for such a view derives from recent investigations of another hallucinatory syndrome, the Charles Bonnet syndrome (Schultz & Melzack, 1991). This syndrome (CBS) occurs primarily in otherwise normal elderly persons, in conjunction with loss of vision due to conditions such as cataracts, glaucoma, or macular degeneration. Under such conditions, up to 30% of patients report experiencing complex and involuntary visual hallucinations while otherwise fully and normally conscious. The real incidence may be substantially higher, since patients are understandably chary of reporting bizarre experiences to medical personnel. These hallucinations typically appear to their subjects as occupying external space, with visual quality higher than that of their residual vision. The precipitating circumstance, in this as in other sensory-deprivation syndromes including dreaming, is not so much the presence of an abnormal activity pattern (as in migraine aura) as the absence of normal input-driven ones.
The especially interesting finding from recent EEG and fMRI imaging studies is that the self-reported phenomenology of hallucinations in CBS patients correlates with increased activation in the corresponding, functionally specialized subregions of (ventral) visual cortex (ffytche, Howard, Brammer, David, & Williams 1998; Santhouse, Howard, & ffytche, 2000). Thus for example if patients hallucinated in color, increased activation was found in the color area (V4). This finding is early and tentative, but if correct it poses problems for a conventional causal account. Clearly, some sort of subliminal process must be organizing and driving these complex, widespread, and coordinated activation patterns, because they cannot plausibly be expected to arise by chance from the spontaneous background activity of the input-deprived visual cortex itself. Analogous though less clear-cut activation patterns are found in the case of voluntary imagery, when subjects deliberately attempt to visualize some pre-specified state of affairs (Kosslyn, 1994), but in CBS, as in dreams and waking apparitions, the imagery is completely unexpected, surprising, and involuntary. Might this not reflect the operation of something very like Myers’s subliminal consciousness, establishing appropriate patterns of brain activity in the process of delivering its products to the supraliminal consciousness? Such a model might also explain the further result of ffytche et al. (1998), unexpected by them, that their fMRI signals arose before emergence of the corresponding qualitative features into subjects’ hallucinatory experiences.
A conventional global-workspace theorist might respond by agreeing that a subliminally formed template or plan of some sort drives the hallucinatory process, while insisting that it be unconscious and inhabit some other part of the brain. This move, however, would not accommodate at least two additional lines of evidence that support a more Myers-like view. First, as already pointed out in Chapter 5, psychological automatisms in the form of involuntary visual imagery and hallucinatory states such as dreaming are known in general to be especially fruitful sources of strong psi effects, showing that the hallucinatory process has access to information beyond the reach of the brain itself. A direct connection to CBS can be made by way of Gurney et al. (1886, vol. 1, pp. 389–456), who discussed the hypnagogic state at length, noting that the number of veridical apparitions occurring under these conditions in their sample of 701 spontaneous cases was far higher than would be expected from the sheer proportion of time we humans pass in such states. The fertility of their analysis has been abundantly confirmed in subsequent work, including later studies of spontaneous cases (Sherwood, 2002) and contemporary experimental studies using the “Ganzfeld” sensory-deprivation technique and other means of inducing hypnagogic-like states.12 Given this background, in fact, we will venture two further predictions in relation to CBS itself: First, inasmuch as CBS amounts to a long-term though partial Ganzfeld condition, we expect that sympathetic and open-minded investigation will reveal it to be a fertile source of spontaneous psi effects as well, potentially adaptable for experimental purposes. Second, just as “shell-hearing,” like crystal-gazing or “scrying,” can elicit psi-conducive forms of hallucinatory activity in normal persons (E. F. Kelly & Locke, 1981a), there ought to be an auditory analog of CBS with similarly psi-conducive properties.
The second line of evidence is of course that developed in detail in our previous chapter. To recapitulate, we showed there that full-blown or even abnormally heightened conscious awareness can occur under conditions of general anesthesia and/or cardiac arrest, conditions in which the specific neurophysiological mechanisms believed by current mainstream neuroscience to be necessary for consciousness have been rendered inoperative. This argument, if correct, shows even more clearly than the argument from psi phenomena that the roots of the hallucinatory process—Myers’s “indwelling general perceptive power”—lie outside the material brain as conventionally understood.
In summary of this section on hallucinatory syndromes, Herschel and Myers now appear to have been rather too hasty in thinking that Herschel’s “geometrical spectres” provided a direct, unmediated window into the operations of the subliminal mind, but in the end they may have been more right than wrong. This topic occupies just over a page in Myers’s own chapter, and it is not particularly essential to his main argument since he could certainly have employed other relevant examples such as post-hypnotic visual hallucinations. We have dwelt upon it at length, however, because these recent empirical developments have more clearly exposed and substantiated the deep connections among the various forms of hallucinatory activity, and their compatibility with Myers’s conception of a fundamental subliminal power or faculty of imagination that enters into the highest expressions of genius.13
Myers turns next to his main topic, the natural history of inspiration itself—”the records, namely, left by eminent men as to the element of subconscious mentation, which was involved in their best work” (HP, vol. 1, p. 89). We will recapitulate the main threads of his exposition, departing for reasons of expository convenience from strictly textual order and emphasizing points of special theoretical significance to Myers. As we proceed, we will also attempt to show how his overall picture of genius has been corroborated and extended through various lines of subsequent research and scholarship.
Myers begins by briefly quoting 10 typical illustrations of the automatic character of inspiration, selecting these from a much larger number of records obtained by Paul Chabaneix, a French physician, through direct inquiry with eminent contemporary artists, philosophers, and writers. Two points are immediately noteworthy here: First, Myers explicitly declined to make use of similar material that had been collected on a much larger scale, but in a far less disciplined manner, by von Hartmann in his then-popular book, Philosophy of the Unconscious, a work which Myers bluntly characterizes as “to me especially distasteful, as containing what seems to me the loose and extravagant parody of important truth” (HP, vol. 1, p. 89). Second, and more importantly, the basic phenomenological picture that Myers seeks to convey here has been confirmed and amplified in many subsequent collections.14
Ghiselin’s introductory essay provides a particularly thoughtful, insightful, and concise descriptive account of the creative process, based on first-person reports of creative activity in a wide variety of fields. His account is also highly consistent with Myers’s as far as it goes, though lacking Myers’s more richly elaborated conception of subliminal operations, and was arrived at independently, making the confirmation of Myers’s general outlook even more striking. In regard to automatism, Ghiselin (1952) could hardly be more definite:
[There is] a sense of self-surrender to an inward necessity inherent in something larger than the ego and taking precedence over the established order…. Production by a process of purely conscious calculation seems never to occur…. More or less of such automatism is reported by nearly every worker who has much to say about his processes, and no creative process has been demonstrated to be wholly free from it. (pp. 15–16)
Inspiration is essentially the intrusion into supraliminal consciousness of some novel form of order that has gestated somewhere beyond its customary margins. The content of such inspirations can vary widely in character, scope, and completeness, but psychologically the process is fundamentally the same throughout its range. The novelty appears
sometimes in the form of a mere glimpse serving as a clue, or like a germ to be developed; sometimes a fragment of the whole, whether rudimentary and requiring to be worked into shape or already in its final form; sometimes essentially complete, though needing expansion, verification, or the like. Spontaneous appearance of inventions very fully formed is not extremely rare, but it is by no means ordinary. Spontaneity is common, but what is given is usually far from complete, (p. 15)
Thus for Ghiselin as for Myers the essence of genius is unusually effective cooperation between supraliminal and subliminal elements of the personality. Neither is sufficient by itself to support the highest forms of creative achievement. Skilled and persistent voluntary effort is almost invariably necessary, both to initiate and shape the subliminal work and to evaluate its products, the materials delivered to waking consciousness in moments of inspiration. But persons of genius throughout history have testified consistently that such moments were the characteristic and essential accompaniment of their best work.
These moments often reach extreme levels of phenomenological impressiveness. Material may suddenly appear that is surprising, unfamiliar, even strange, flowing with extraordinary ease and copiousness, accompanied by intense affect and excitement, and in the absence of any feeling of personal responsibility for what comes. Many examples of such events can be found in the sources cited above, and we need not repeat them in profusion here, but it is precisely in light of such extreme and unusual manifestations that Myers seeks to establish the fundamental kinship between subliminal uprushes of genius and other forms of psychological automatism.
Sleep and dreams provide a first sort of example, relatively close to ordinary experience. Myers pictures the genius as successfully coordinating the waking and sleeping phases of his existence: “He is carrying into sleep the knowledge and the purpose of waking hours;—and he is carrying back into waking hours again the benefit of those profound assimilations which are the privilege of sleep” (HP, vol. 1, p. 90). Robert Louis Stevenson provides the chief example here: In his well-known chapter on dreams in Across the Plains, Stevenson describes how he progressively harnessed his abundant dream-life in service of his writing. His “Brownies,” the “little people” who manage his internal theater,
are near connections of the dreamer’s, beyond doubt; they share in his financial worries and have an eye to the bank-book; they share plainly in his training;…they have plainly learned like him to build the scheme of a considerate story and to arrange emotion in progressive order; only I think they have more talent; and one thing is beyond doubt;—they can tell him a story piece by piece, like a serial, and keep him all the while in ignorance of where they aim. (HP, vol. 1, p. 91)
The Brownies are thus capable of “really guileful craftsmanship,” and though occasionally guilty of absurdities they generally fashion “better tales than he could fashion for himself.” Stevenson concludes that his unseen collaborators do the bulk of his creative work: “That part which is done while I am sleeping is the Brownies’ part beyond contention; but that which is done when I am up and about is by no means necessarily mine, since all goes to show the Brownies have a hand in it even then” (HP, vol. 1, p. 91).
In his following chapter on sleep (section 417 and its appendix), Myers notes several additional instances of nocturnal problem-solving, which in that chapter provide him with stepping-stones toward dreams more clearly involving acquisition of supernormal knowledge. In three of these instances the solution emerged during a dream, while in the fourth (Professor Lamberton) it appeared upon awakening. The last is in some ways the most interesting: Lamberton had tried in vain for over a week to solve algebraically a certain problem in analytic geometry, but upon awakening that morning he found the solution standing before his eyes in the form of a complex hallucinatory diagram drawn upon the opposite wall, formerly a blackboard. This impressed him greatly, as he had made no previous efforts to solve the problem geometrically, was in general a poor visualizer, and had no other waking hallucinatory experiences either before or after this one.
Mazzarello (2000) summarizes the well-known cases of scientific discovery involving Kekulè, Loewi, and Mendeleyev, and many other cases of creative solutions obtained during sleep, dreams, and hypnagogic states have been collected by D. Barrett (2001), Mavromatis (1987), and Van de Castle (1994), among others. We will cite one more, which would have been especially welcomed by Myers as a demonstration of subliminal contributions to higher mathematics. The distinguished mathematician Hadamard (1949) states that:
One phenomenon is certain and I can vouch for its absolute certainty: the sudden and immediate appearance of a solution at the very moment of sudden awakening. On being very abruptly awakened by an external noise, a solution long searched for appeared to me at once without the slightest instant of reflection on my part—the fact was remarkable enough to have struck me unforgettably—and in a quite different direction from any of those which I had previously tried to follow, (p. 8)15
More generally, drawing not only upon his own experiences of mathematical discovery but also upon extensive and direct interactions with many other leading mathematicians and scientists, Hadamard declares that any doubt as to the existence of unconscious work “can hardly arise” (p. 21). Indeed, “strictly speaking, there is hardly any completely logical discovery. Some intervention of intuition issuing from the unconscious is necessary at least to initiate the logical work” (p. 112).
We jump now to section 327 of HP, in which Myers begins to develop more fully the deep and overlapping interconnections among genius, automatism, and trance. All, of course, involve subliminal influences upon the supraliminal consciousness, but these can vary greatly in terms of their intensity, the degree to which they alter the normal functioning of that consciousness. Just as trance, when habitual, tends to engender automatic writing or speech, and prolonged automatism tends to induce trance, the subliminal uprushes of genius can be arranged in a hierarchy of increasing involvement of these more extreme manifestations.
At the near or shallow end of this progression in depth is the momentary flash of inspiration, a brief automatism. Myers illustrates this with lines from Wordsworth describing how “Some lovely image in the song rose up / Full-formed, like Venus rising from the sea,” adding that “such a sudden poetic creation, like the calculating boy’s announcement of the product of two numbers, resembles the sudden rush of planchette or pencil, in haste to scrawl some long-wished-for word” (HP, vol. 1, p. 104).
A transition from ordinary facility to something involving more significant contributions from the subliminal can sometimes be recognized in connection with improvisation. Myers first acknowledges the important contributions of memory and conventional rules or habits to ordinary musical or oratorical improvisation, following lines quite similar to those subsequently taken by Boden (1991) in regard to computer simulations of jazz improvisation and the like. But in George Sand ordinary facility evolved into
an unusual vigour and fertility of literary outflow going on in an almost dreamlike condition; a condition midway between the actual inventive dreams of R. L. Stevenson and the conscious labour of another man’s composition… a continuous and effortless flow of ideas, sometimes with and sometimes without an apparent externalisation of the characters who spoke in her romances. (HP, vol. 1, p. 106)
Similarly, Charles Dickens was highly prone to hypnagogic-like reveries (Mavromatis, 1987) and alluded often to the tendency of his imaginary characters toward “independence”; for example, “Mrs. Gamp, his greatest creation, spoke to him, he tells us (generally in church), as with an inward monitory voice” (HP, vol. 1, p. 106).
Myers next describes the even more extreme case of M. de Curel, a distinguished French dramatist, of which a long account had been published by the psychologist Alfred Binet in 1894. In de Curel, Myers observes, the waking experience of creation approaches that of Stevenson in dreams:
He begins in an ordinary way, or with even more than the usual degree of difficulty and distress in getting into his subject. Then gradually he begins to feel the creation of a number of quasi-personalities within him;—the characters of his play, who speak to him;—exactly as Dickens used to describe Mrs. Gamp as speaking to him in church. These personages are not clearly visible, but they seem to move round him in a scene—say a house and garden—which he also dimly perceives, somewhat as we perceive the scene of a dream. He now no longer has the feeling of composition, of creation, but merely of literary revision; the personages speak and act for themselves, and even if he is interrupted while writing, or when he is asleep at night, the play continues to compose itself in his head. Sometimes while out shooting, &c, and not thinking of the play, he hears sentences rising within him which belong to a part of this play which he has not yet reached. He believes that subliminally the piece has been worked out to that further point already. M. de Curel calls these minor duplications of personality a bourgeonnement or budding of his primary personality;—into which they gradually, though not without some painful struggle, re-enter after the play is finished. (HP, vol. 1, p. 107)
Myers goes on at this point to other topics to which we will return shortly, but we will first point out just a few of the many additional examples that can be found in collections like those cited earlier, in which works of genius are mediated at least in part by automatism, sometimes accompanied by trance-like altered states of consciousness.
Parallels to the literary improvisation of George Sand, for example, can be found in dance, in the work of choreographer George Balanchine:
He started doing this movement and that, showing the dancers what they had to do. Then at a certain moment it became something much more than just himself and his ideas. He started to work as a somnambulist, without knowing what he was doing. And all this was quickly done, with the greatest assurance. When he finished, he would sit and ask the dancers to show him what he had done, and he would seem to be very astonished. That is what I call inspiration. (Teachout, 2004, pp. 18–19)
A. E. Housman (1952) described how the inspiration for much of his best poetry occurred during afternoon walks, following a pint of beer at lunch. As he strolled along, absent-mindedly looking about,
there would flow into my mind, with sudden and unaccountable emotion, sometimes a line or two of verse, sometimes a whole stanza at once, accompanied, not preceded, by a vague notion of the poem which they were destined to form part of. Then there would usually be a lull of an hour or so, then perhaps the spring would bubble up again…. I happen to remember distinctly the genesis of the piece which stands last in my first volume. Two of the stanzas, I do not say which, came into my head, just as they are printed, while I was crossing the corner of Hampstead Heath between Spaniard’s Inn and the footpath to Temple Fortune. A third stanza came with a little coaxing after tea. One more was needed, but it did not come: I had to turn to and compose it myself, and that was a laborious business. I wrote it thirteen times, and it was more than a twelvemonth before I got it right, (p. 91)
William Blake, who with Wallace Stevens was “an extremist of the imagination” (Brann, 1991, p. 509), wrote his friend Butts that his prophetic poem Milton was written “from immediate dictation, twelve or sometimes twenty or thirty lines at a time, without premeditation and even against my will. The time it has taken in writing was thus rendered non-existent, and an immense poem exists which seems to be the labour of a long life, all produced without labour or study…. I may praise it, since I dare not pretend to be other than the secretary” (Damon, 1958, p. 202).16
The hypermnesic, possibly eidetic, and mammothly prolific writer Thomas Wolfe (1952) describes the onset of the process that generated three enormous novels in 4½ years:
An extraordinary image remains to me from that year, the year I spent abroad when the material of these books first began to take on articulate form. It seemed that I had inside me, swelling and gathering all the time, a huge black cloud, and that this cloud was loaded with electricity, pregnant, crested, with a kind of hurricane violence that could not be held in check much longer; that the moment was approaching fast when it must break. Well, all I can say is that the storm did break. It broke that summer while I was in Switzerland. It came in torrents, and it is not over yet.
I cannot really say the book was written. It was something that took hold of me and possessed me, and before I was done with it—that is, before I finally emerged with the first completed part—it seemed to me that it had done for me. It was exactly as if this great black storm cloud I have spoken of had opened up and, mid flashes of lightning, was pouring from its depth a torrential and ungovernable flood. Upon that flood everything was swept and borne along as by a great river. And I was borne along with it. (p. 187)
Nietzsche (1952) recounts in similarly extreme terms in Ecce Homo how he was “invaded” by Thus Spake Zarathustra:
Can any one at the end of this nineteenth century possibly have any distinct notion of what poets of a more vigourous period meant by inspiration? If not, I should like to describe it. Provided one has the slightest remnant of superstition left, one can hardly reject completely the idea that one is the mere incarnation, or mouthpiece, or medium of some almighty power. The notion of revelation describes the condition quite simply; by which I mean that something profoundly convulsive and disturbing suddenly becomes visible and audible with indescribable definiteness and exactness. One hears—one does not seek; one takes—one does not ask who gives: a thought flashes out like lightning, inevitably without hesitation—I have never had any choice about it. There is an ecstasy whose terrific tension is sometimes released by a flood of tears, during which one’s progress varies from involuntary impetuosity to involuntary slowness. There is the feeling that one is utterly out of hand, with the most distinct consciousness of an infinitude of shuddering thrills that pass through one from head to foot;—there is a profound happiness in which the most painful and gloomy feelings are not discordant in effect, but are required as necessary colors in this overflow of light. There is an instinct for rhythmic relations which embraces an entire world of forms (length, the need for a widely extended rhythm, is almost a measure of the force of inspiration, a sort of counterpart to its pressure and tension). Everything occurs quite without volition, as if in an eruption of freedom, independence, power and divinity. The spontaneity of the images and similes is most remarkable; one loses all perception of what is imagery and simile; everything offers itself as the most immediate, exact, and simple means of expression, (pp. 202–203)
Harding (1948, chap. 2) cites a large number of additional cases in which the sense of being possessed by some external force and compelled to create is especially pronounced. One involves Goethe, who described how his poems sometimes
have come suddenly upon me, and have insisted on being composed immediately, so that I have felt an instinctive and dreamy impulse to write them down on the spot. In such a somnambulistic condition, it has often happened that I have had a sheet of paper lying before me all aslant, and I have not discovered it till all has been written, or I have found no room to write any more. I have possessed many such sheets written diagonally, (p. 12)
Automatism also figures prominently in an especially curious case involving the great 20th-century poet James Merrill, and in particular his masterpiece, The Changing Light of Sandover (1980/1993). On the cover of this book Harold Bloom expresses particular admiration for one of its chapters: “I don’t know that The Book of Ephraim, at least after some dozen readings, can be over-praised, as nothing since [by] the greatest writers of our century equals it in demonic force.” Yet the book was produced in its entirety by means of a ouija board operated jointly by Merrill and his longtime friend “R.” In its opening pages Merrill describes the instant the ouija board came to life: “YES a new and urgent power YES / Seized the cup. It swerved, clung, hesitated, / Darted off, a devil’s darning needle / Gyroscope our fingers rode bareback / (But stopping dead the instant one lost touch).” Thus if either Merrill or his partner took his finger off the cup, it stopped dead. But when it worked, it was astonishing: “Yet even the most fragmentary message—/ Twice as entertaining, twice as wise / As either of its mediums—enthralled them.” It is of course difficult if not impossible to determine exactly who is contributing what to the creative process in a case of this sort, but the suggestion that the results outstrip what either could accomplish alone is especially intriguing and worthy of further investigation.17
Bowra (1955) summarizes in more general terms the involuntary character of inspiration as experienced and reported by many poets:
[It] manifests itself in a manner which no one can mistake. The poet unaccountably finds himself dominated by something which absorbs his being and excludes other interests from his mind. It is not easy to define exactly what this is, but we may mark certain elements in it. Central to it is something which may be called an idea, though in some ways it is too vague to deserve the name. It has a powerful character and atmosphere of its own, and though at first it is too indefinite for intellectual analysis, it imposes itself on the poet with the majesty and authority of vision. Even if he does not fully understand it, he feels it and almost sees it. (p. 4)
In sum, Myers seems to us certainly correct in pointing out connections of genius with trance and automatism. Indeed, these same connections can also be recognized from the opposite direction—that is, from cases in which the aspects of trance and automatism predominate, but which also give rise to high-level creative production. We turn next to these.
In the first instance it should be noted that considerable amounts of “creativity” are necessarily involved in the production of characteristic dissociative phenomena such as glove anesthesias, negative hallucinations in hypnosis, and “alter” or “multiple” personalities formed in response to overwhelming trauma. Philosopher Stephen Braude (2002) has shown with particular clarity that such phenomena cannot be conceived in terms of atomic mental states or contents that are statically segregated by fixed and passive boundaries of some sort. Rather, successful maintenance of the dissociation requires continuous, active, adaptive—in short, creative—subliminal improvisation in response to the subject’s ongoing and constantly-changing interactions with his environment. Mediumistic dramatizations of deceased personalities also sometimes display (at minimum) remarkable histrionic capacities, reproducing voice, tone, mannerisms, characteristic turns of phrase, and other traits of the deceased with verisimilitude sufficient to convince even knowledgeable and critical sitters of their continuing post-mortem existence.
Cases involving emergence of secondary personalities with characteristics superior to those of the primary personality are also relevant here. One such case is that of Old Stump, described in Chapter 5. Another is de Puységur’s patient Victor Race, suffering from an inflamed lung, who manifested when mesmerized a secondary personality remarkably more gifted than Victor in his normal state: “Though ordinarily a simple and tongue-tied peasant, he would, in the somnambulic state, converse in a fluent and elevated manner…. Victor assumed management of his own case, diagnosing and prescribing for his illness, and predicting its course. More than this: on being brought into contact with other patients, he seemed able to do the same for them” (Gauld, 1992, p. 41).
Still more important is the well-studied case of the Swiss medium “Hélène Smith” (Catherine Elise Miiller). Myers (HP, vol. 2, pp. 130–144) provides a précis based on the initial investigation by Flournoy (1900/1963), using this as his primary example of the action of the mid-level “mythopœic” subliminal centers that engage perpetually in a “strange manufacture of inward romances” (an interpretation approved by Ellenberger, 1970, pp. 150, 317–318). The initial stages of Hélène Smith’s trance mediumship were replete with automatisms of various kinds but consisted mainly of lengthy descriptions of her supposed adventures in a series of increasingly exotic settings, including India and the planet Mars. Some of these reports seemed to contain considerable amounts of plausible historical and geographical detail, encouraging her admirers to agree with her that she was recounting her own actual experiences of previous lives. Flournoy and Myers, however, saw these tales rather as moderately creative productions of her subliminal, drawing upon a faculty of “cryptomnesia” (Stevenson, 1983a), or subliminal access to materials long ago encountered but inaccessible to ordinary conscious recall. Flournoy’s book recounts in detail how he was able by laborious investigation to track down probable sources for much of the content of her romances, although he did in the end conclude that they also contained at least a few elements of probable supernormal origin (Flournoy, 1911).
The story does not end there, however. Hélène Smith’s mediumship continued to evolve, eventually supplanting her previous waking life as a department manager in a large shop in Geneva. Flournoy was unfortunately not privy to these later developments, as she and her followers had become irritated with his deflationary zeal and punished him by refusing further cooperation. Of particular interest is a large volume of automatic drawings and paintings, many of which are reproduced in a massive study by Deonna (1932). Even in reproductions without color her paintings are well composed, smoothly executed with defined images, exuding a surreal religiosity comparing favorably with the paintings of Frida Khalo. Yet Hélène Smith apparently painted without a brush, using her fingers in a trance-like condition. N. Fodor (1966) quotes her own description of the process:
On the days when I am to paint I am always roused very early—generally between five and six in the morning—by three loud knocks at my bed. I open my eyes and see my bedroom brightly illuminated, and immediately understand that I have to stand up and work. I dress myself in the beautiful iridescent light, and wait a few moments, sitting in my armchair, until the feeling comes that I have to work. It never delays. All at once I stand up and walk to the picture. When about two steps before it I feel a strange sensation, and probably fall asleep at the same moment. I know, later on, that I must have slept because I notice that my ringers are covered with different colours, and I do not remember at all to have used them. (p. 350)18
A related case of perhaps even greater theoretical significance is that of “Patience Worth,” first studied and reported in depth by Walter Franklin Prince (1927/1964). The medium in this case was Pearl Curran, a midwestern American housewife of modest education who possessed, so far as Prince could determine through extensive and careful investigation, no special interests or aptitude in language, literature, or poetry. In Pearl’s 31st year, in the course of some half-hearted experimentation with a ouija board, the personality identifying itself as Patience Worth, ostensibly a 17th-century Englishwoman, made its first appearance. Her ensuing collaboration with Pearl lasted almost 25 years, and ultimately produced some 29 volumes or 4,375 single-spaced pages of novels, poems, proverbs, aphorisms, witticisms, and conversational repartee (Braude, 2003).
Far more important than the sheer volume of this material, however, are several aspects of its character. First, it was generated automatically, initially by means of the ouija board, later by automatic vocalization of single letters, then words, apparently accompanied by panoramas of vivid visual imagery illustrating or dramatizing the spoken content. Pearl occasionally carried out directly competing tasks such as writing a letter to a friend while this was going on. Second is the extreme fluency and virtuosity of the performance. Patience typically produced her material without delay and faster than a stenographer could write it down, pausing only to let the stenographer catch up. This was the case even in response to difficult challenges posed on the spot by Prince—for example, to improvise a poem on a given theme, or to write alternating lines or passages on different subjects or in different literary styles. Interrupted by Prince in the middle of a sentence or passage, Patience could resume on demand and without break or pause, sometimes weeks or months later. Her material was consistently of high quality even as initially produced, and she almost never revised anything. She also displayed extensive and accurate knowledge of the various times and places figuring in her novels, and portrayed them with impressive verisimilitude. Perhaps most remarkable of all, she possessed an extensive repertoire of archaic Anglo-Saxon words deriving from multiple periods and localities, many of which were known at the time, if at all, only to a few professional philologists.
In short, Patience Worth, who had emerged essentially fully-formed in those first sessions with the ouija board, differed radically from the relatively staid and ordinary Pearl Curran, and not only in matters of personality, style, and attitude but also in knowledge, intelligence, and creative capacities. Her performances, as Prince (1927/1964) points out, required “phenomenal memory, phenomenal speed, and phenomenal complexity of mental operations” (p. 487). Summing up the results of his “arduous and unremitting labor,” Prince (1927/1964) advances the following proposition; “Either our concept of what we call the subconscious must be radically altered, so as to include potencies of which we hitherto have had no knowledge, or else some cause operating through but not originating in the subconsciousness of Mrs. Curran must be acknowledged” (p. 509). We can only second this theoretical challenge, as does Schiller (1928), as well as Braude (2003) in his valuable summary and analysis of the case.
Additional examples of the connection between mediumship and genius are presented and discussed in Grosso (in press). One we cannot resist mentioning here, to conclude this section, concerns the role of Myers and automatism in Surrealism, which was perhaps the most important artistic movement of the 20th century. Cubism, Dada, Futurism, Abstract Expressionism, and Minimalism were all important but limited innovations, primarily technical and thematic. Surrealism alone challenged fundamental premises about art and creativity, shifting the focus from conscious to unconscious processes, introducing the role of chance in the creative process, and treating that process as not merely aesthetic but political, social, and metaphysical. Although it is widely supposed that Surrealism was inspired wholly by Freud, that is certainly not correct: Its chief theoretician, Andre Breton, published in 1933 an article specifically acknowledging its indebtedness to “the gothic psychiatry of F. W. H. Myers” (Breton, 1933/1997). Myers’s work on automatism in fact provided the key psychological mechanism that Surrealism would attempt to exploit in novel ways: “Surrealism has above all worked to bring inspiration back into favor, and we have for that purpose promoted the use of automatic forms of expression” (p. 15). The goal of Surrealism is essentially to unify the personality, which means for Breton what Myers meant by genius, the successful coordination and interpenetration of dream and waking life. As Breton (1924/1972) puts it in his Manifestoes of Surrealism: “I believe in the future resolution of these two states, dream and reality, which are seemingly so contradictory, into a kind of absolute reality, a surreality, if one may so speak” (p. 14). It would be an interesting study, but one which would take us too far afield for present purposes, to explore in greater detail the role actually played by psychological automatisms in the production of particular works of Surrealist art.
To summarize the discussion so far: Myers argues, and we think he is correct, that whereas momentary “flashes of inspiration” are essentially brief automatisms, more or less continuous with much of what goes on in ordinary mental life, in more extreme cases of real genius the subliminal uprush becomes more intense and protracted, and arises from deeper subliminal strata that utilize different modes of operation and may have access to additional sources of information. This leads to the third major feature of his analysis of inspiration, and—surprisingly—to a cluster of issues lying at the heart of contemporary cognitive theory.
Myers introduces this theme in section 322 (HP, vol. 1, p. 98).
And thus there may really be something at times incommensurable between the inspirations of genius and the results of conscious logical thought. Just as the calculating boy solves his problems by methods which differ from the methods of the trained mathematician, so in artistic matters also that “something of strangeness” which is in “all excellent beauty”, may be the expression of a real difference between subliminal and supraliminal modes of perception.
In the next six pages, one of the most condensed and difficult but important parts of the chapter, Myers develops his central ideas: Subliminal mentation is less closely bound than supraliminal mentation to language, either ordinary spoken and written language or the specialized languages of science and mathematics; but it is not for that reason to be presumed inferior, that it “in some way falls short of the standard implied in articulate speech” (HP, vol. 1, p. 98).
Myers clearly acknowledges the primacy of language as the privileged means of ordinary communication. It is an absolute necessity of intellectual life and the foundation of all civilization. But it is not the whole story: “There is, however, no a priori ground for supposing that language will have the power to express all the thoughts and emotions of man” (HP, vol. 1, p. 99). Indeed, the study of automatisms in general and the inspirations of genius in particular reveal that other forms of symbolism become increasingly important as we access deeper strata of the subliminal. There is a “hidden habit of wider symbolism, of self-communion beyond the limits of speech” that is better adapted to expression of “that pre-existent but hidden concordance between visible and invisible things, between matter and thought, between thought and emotion, which the plastic arts, and music, and poetry do each in their own special field discover and manifest for human wisdom and joy” (HP, vol. 1, p. 101).
Myers goes on in section 326 to make some extremely condensed and penetrating remarks elaborating this basic conception with respect to poetry, music, and the plastic arts in turn. We encourage interested readers to savor thoughtfully these remarks, which anticipate in a remarkable way attempts by later philosophers such as Brann (1991), Cassirer (1955–1996), Langer (1956), and Whitehead (1938/1968) to temper the linguistic obsessions of modern analytic philosophy with an appreciation of non-discursive or “presentational” modes of symbolism. Langer (1956) says of music, for example, that “there are certain aspects of the so-called ‘inner life’—physical or mental—which have formal properties similar to those of music—patterns of motion and rest, of tension and release, or agreement and disagreement, proportion, fulfillment, excitation, sudden change, etc.” (p. 185). By virtue of such isomorphisms a musical configuration can become expressive of an emotional state, so that “music sounds the way emotions feel.”
But let us return now to mainstream cognitive psychology, where a similar imbalance in favor of discursive and propositional forms of thought has long dominated the scene, and especially since the rise of the computational theory of the mind (CTM; Chapter 1). We wish first simply to reiterate Myers’s central empirical claim, that subliminal contributions to creative activity generally, and to the more extreme “inspirations of genius” in particular, tend strongly to take the form of unusually intense and often involuntary imagery, predominantly though not exclusively visual, accompanied by strong affect and a profound sense of surprise and wonder at its beauty, harmony, elegance, and proportion or fitness to the consciously intended purpose. That this claim is broadly correct, descriptively, seems to us beyond reasonable doubt, as we have already indicated. What we want next to discuss in greater depth concerns rather the manner in which scientific psychology subsequent to Myers has dealt with such matters.
The CTM itself in its classical symbol-processing form scarcely touches these subjects at all, with one important exception—analogy—that we will discuss in more detail below. Most of the remaining work, surveyed by Boden (1991; updated in Boden, 1999), involves attempted computer simulations of processes of scientific discovery, musical improvisation, or the creation of simple stories, poems, and the like. The authors of these various scientific-discovery programs have been widely faulted for setting things up in advance in such a way as to make their “discoveries” inevitable, a criticism with which we generally agree, and in that light the overall assessment of this work by Boden herself, generally an uncritical enthusiast of the CTM, is telling: “The success of programs for scientific discovery, limited though it is, is not matched by programs for producing poems or (especially) stories” (Boden, 1999, p. 359).
We will leave this part of the subject there, adding only that the reasons for these failures, in our view, are essentially the central defects of the CTM as identified by critics such as Searle and Dreyfus and summarized in Chapter 1. It is also worth noting, however, that Hofstadter and FARG (1995) arrive at an equally negative overall assessment from the opposite, CTM-friendly, point of view. It should also be pointed out, we think, that all such attempts to explain genius as a completely rule-bound process conflict with a deep intuition, shared by many, that genius by definition breaks old rules and makes new ones: As Kant (1790/1951) put the point so clearly, “genius is a talent for producing that for which no definite rule can be given; it is not a mere aptitude for what can be learned by a rule” (p. 150). Indeed, he says, “Genius is the talent (or natural gift) which gives the rule to art.”
By far the largest proportion of attempts to explain the creative process have invoked the core doctrines of “associationism,” in forms ranging from the 17th-century mechanical theory of association of ideas to its modern descendant, connectionism, which speaks in terms of waves of excitation reverberating through neural networks (Chapter 1). (As J. Fodor [2001] remarks, perhaps a bit too caustically: “That’s what’s so nice about empiricist cognitive science: You can drop out for a couple of centuries and not miss a thing” [p. 49].) Wallas (1926), for example, thought of unconscious or preconscious ideas as existing ready-made in a Jamesian-style “fringe” of consciousness, in effect a remote part of the “upper cortex,” which he and others of that period naively conceived as a kind of vast telephone switchboard. The creative person is one in whom trains of association set in motion by the preparatory labor somehow penetrate that remote region during the incubation phase and deliver useful ideas back to consciousness in the form of inspiration. Ghiselin (1952) expresses very similar views in a slightly updated neurophysiological terminology.
Later accounts by psychologists followed similar lines but sought to impose a more mechanistic flavor consistent with the observations of Guilford (1950) and others regarding the importance of “divergent thinking” and the like. So, for example, Mednick (1962) advanced an associationist theory according to which creativity is found in persons and conditions in which the associative hierarchies linking concepts are relatively equipotential or “flat,” permitting large numbers of remote and surprising associations, in contrast with “steep” hierarchies that permit only relatively small numbers of obvious or stereotypical ones. Mendelsohn (1976) related these characteristics to attentional processes, emphasizing that “defocusing” of attention increases the range of possible associations and combinations of ideas and hence, it is presumed, creativity.
Martindale (1995, 1999) has recently updated these ideas within the framework of contemporary neural-network and connectionist theorizing, seeking in effect a neurobiology of the creative process. He reviews, for example, a variety of evidence suggesting that the flat associative hierarchies and defocused attention that supposedly account for creativity correlate to some degree with brain conditions such as reduced cortical arousal and reduced frontal-lobe control, (possibly) accompanied by relatively greater right-hemisphere involvement.19 The relentless psychometrician Hans Eysenck (1995) welcomes these “state” formulations, treating them as complementary to his own “trait” concept of “psychoticism” as a biologically grounded propensity toward remote but relevant associations (p. 267). For both Martindale and Eysenck, however, as for many other leading creativity theorists, the central notion that “remote associations” comprise the essence of the creative process goes essentially unchallenged.
The fundamental idea of associationist theories of creativity is that the incubation/inspiration cycle can satisfactorily be conceptualized in terms of retrieval and combination of ideas already in the network, by means of some process of spreading activation. This conception seems plausible enough, we would agree, in connection with certain rudimentary forms of the phenomena to be explained. It might well account, for example, for what goes on when we finally retrieve a temporarily forgotten word following unsuccessful attempts to do so in the tip-of-the-tongue (TOT) state (A. S. Brown, 1991; R. Brown & McNeill, 1966; James, 1890b, vol. 1, p. 251). As many observers have noted, such occasions seem to involve—at least in germ—all four stages of the Wallas model, so why not regard TOT as a suitable prototype for the whole domain?
Forceful answers can be found, we submit, in a source rarely consulted by contemporary psychologists, namely the investigations by 19th-century English romantic poets, Coleridge in particular, into the nature of human imagination.20 The gist of the story goes as follows: Coleridge had felt a strange power in certain early poems by his friend Wordsworth, poems that he recognized as markedly superior to his own. Seeking to understand this puzzling but important effect, and thus to harness it in service of his own poetry, he turned first to the mainstream psychological theory of his day, the theory of association of ideas as developed by figures such as Hobbes, Locke, Hume, and Hartley. This theory represented a deliberate and self-conscious attempt to extend the triumphs of classical mechanistic physics into the domain of the mind. The “atoms” of the new science were “ideas,” typically conceived as fainter replicas of former perceptions, or as the simpler qualities or parts into which such perceptions could be analyzed. Their commerce was to be governed by “laws of association” analogous to Newton’s laws of motion, based in particular on factors such as contiguity in time or place, resemblances of one or another sort, and connections of cause and effect. Imagination, in this framework, was to be conceived in terms of novel combinations or recombinations of these independent, self-existent atomic parts. Thus the prototypical example of the imagination at work became the Chimera of Greek mythology, a fire-breathing monster which conjoined the head of a lion with a goat’s body and serpent tail. Correspondingly, imagination was regarded as fundamentally delusive, and figurative and metaphorical language were viewed as serving purely ornamental functions.
Coleridge struggled to reconcile this mechanistic theory with what he could directly observe in the workings of real human imagination as manifested in creative geniuses such as Wordsworth, Shakespeare, Milton, and himself, among sundry others. Imagination, he concluded, cannot be fully understood in these terms; moreover, the unexplained residue lies at the very heart of creative imagination, and encompasses its most significant properties. Coleridge sought to explicate these properties in terms of his famous distinction between mere fancy and imagination proper: Fancy seems within reach of associationist theories; mechanical and passive, it associates, aggregates, collates, collocates, juxtaposes, transposes, rearranges, and mirrors, giving rise in such ways to low-grade products like the infamous Chimera. Imagination, by contrast, “that synthetic and magical power,” is for Coleridge something radically different and altogether superior: Imagination is organic and active; it assimilates, dissolves and recreates, fuses, synthesizes, and unifies. It transmutes the chaos of raw materials provided by everyday experience, forging and shaping them by means of its inherent “coadunating,” “alembic,” and “esemplastic” powers into truly novel creations that balance or reconcile seemingly opposite or discordant qualities in harmonious unity. It is above all a unique form of thought, and one of the principal powers of the human mind.
A fuller appreciation of Coleridge’s concept of imagination can best be obtained by studying the numerous examples adduced in the sources cited above. We ourselves have no doubt that it poses severe challenges, not only to classical association theories but to cognitive science more generally, and we think that cognitive science can only be enriched by paying greater attention to these long-neglected but very high-level discussions. Two further observations may serve to reinforce this contention.
The first has to do with the subsequent history of Coleridge’s attempted distinction between fancy and imagination, which has so far been pursued mainly in the context of literary theory and criticism. Most such observers have found Coleridge’s distinction ultimately untenable, but in an interesting and important way: Specifically, what they deny is not that imagination has the sorts of properties that Coleridge ascribes to it, but rather that these difficult properties can be prevented from invading the supposedly lesser realm of fancy. That is (and again in accord with Wind’s principle), even relatively ordinary forms of cognition, when more closely examined, are inescapably saturated with properties of the sort that Coleridge viewed as beyond the reach of any mechanistic association theory (Abrams, 1958; I. A. Richards, 1960).
Second, we wish to comment briefly on the few other recent psychological publications known to us to have made contact with this older literature. The first (Sutton, 1998) attempts to deny that there is any problem. Specifically, Sutton claims that certain secondary strains within classical asso-ciationism, epitomized in particular by Hartley, are the true ancestors of contemporary connectionism, and that these are immune to critiques such as Coleridge’s which supposedly apply only to its less sophisticated forms. In fact, however, Coleridge knew Hartley’s work intimately and used his version of associationism as the principal vehicle for his critical analyses. In our view, Sutton simply avoids coming to grips with the real cognitive issues, appealing to abstract connectionist principles such as “distributed representations” and “superimposition” that in fact do little even to help solve problems of ordinary memory function (Chapter 4), let alone those of creative imagination.
Similar comments apply to Boden’s (1991) chapter on “creative connections,” the setting for which is the remarkable investigation by Lowes (1927) of the workings of Coleridge’s own poetic imagination in the production of Kubla Khan and The Ancient Mariner. Having dutifully acknowledged the properties of strong imagination that both Coleridge and Lowes themselves regarded as truly fundamental, Boden proceeds to treat the subject as if it really were mostly a matter of associative connections, and explainable at least in general terms by the elementary properties (which she describes) of connectionist networks. She forthrightly acknowledges the limitations of current models, and that “the intuitive understanding offered by Livingston Lowes is incalculably richer, and more subtle, than any connectionist explanation” (p. 130), but glibly reassures her readers that these gaps will be closed in due course by the “more powerful” computer systems of the future.
No warrants have as yet been provided for these computationalist promissory notes, and as explained in Chapter 1 we strongly doubt that they can ever be fulfilled. The situation is to us strongly reminiscent of the one that provoked the destructive review by Chomsky (1959) of Skinner’s book on verbal behavior, in that connectionist notions which so far explain rather little when deployed in a careful and technically rigorous manner are being brandished about metaphorically by less scrupulous enthusiasts as the answer to everything. A telling commentary on the real state of things in regard to connectionist modeling of creative imagination is that neither “creativity” nor “imagination,” let alone cognate terms such as “inspiration” or “genius,” even appear in the indexes of connectionist handbooks such as Bechtel and Abrahamson (2002) and Harnish (2002).
Our third example is an important paper by N. J. T. Thomas (1999), who approaches the subject of imagination in the context of a review of the “imagery debates” mentioned in Chapter 1. Thomas recognizes that current imagery theories fall far short of providing an adequate theory of imagination, and that the Romantic-era theorists of the imagination had important things to say—sentiments with which we wholly agree. He goes on, however, to a summary dismissal of Coleridge and his allies on grounds that, although descriptively on target, they left us only “a mixture of warmed-over associationist notions, clearly inadequate to the demands being placed upon them, and fragments of grandiose idealist metaphysics, ripe for mystificatory appropriation” (p. 232). Thomas himself turns instead to “perceptual activity” theory, which takes “seeing-as” as the fundamental capacity of imagination, and seeks to derive imagination from an account of perception inspired mainly by recent work in robotics. To this we would respond first that his investment of such high hopes in robotics seems forlorn; but more importantly, (1) that the Coleridge/Myers view is much more like his own than he appreciates (the entire first half in particular of the statement quoted above is simply inaccurate), and (2) that its empirical and philosophical foundations are far stronger than he realizes, as demonstrated both by HP itself and by the present volume.
Despite these caveats, we welcome Thomas’s paper as one of several recent developments within mainstream psychology itself that indicate growing recognition of the need to deal with creative imagination as it actually is, rather than as most contemporary workers have pictured it to be. Even the hard-core experimentalist George Mandler (1994) was moved to remark that “there may well be conditions in which such spreading or fanning out [of network activation] reaches recesses of the unconscious that dynamic psychology has considered, but experimental psychology has not” (p. 25). Another recent example is the “creative cognition” approach (S. M. Smith, Ward, & Finke, 1995; Ward, Finke, & Smith, 1995; Ward, Smith, & Finke, 1999; Ward, Smith, & Vaid, 1997), which in its emphasis on continuity and the role in creative thinking of “pre-inventive structures” such as imagery, analogy, and symbolism, has also begun moving beyond the current frontiers of mainline cognitive theory in a Myers-like direction, albeit in what seems to us so far an excessively timid and limited way.
In sum, what seems to be in the offing is a revival of psychological interest in creative imagination as a distinctive and powerful mode of thought. Just as “genius” has degenerated in recent times into “creativity,” the original Greek phantasia with its connotations of world-fashioning or cosmogonic power has degenerated over the centuries into mere “fantasy,” escapist day-dreaming of little or no cognitive import. Like Brann (1991), we believe psychology must now reverse this pernicious trend; the “sober romanticism” which she champions in her monumental interdisciplinary compilation of thinking about the imagination is in fact highly consistent with Myers’s views, and would have been greatly enriched by those views had only she known them.
We said “revival” in the previous paragraph because imagination has not always been so neglected by psychology as in recent decades. In particular, and as hinted in the statement above by George Mandler, psycho-dynamic theories associated with the clinical wing of the field have always paid far more attention to this subject. In their terms, the “incommensurability” that Myers pointed out between subliminal and supraliminal modes of functioning maps rather nicely onto a famous distinction between “primary” and “secondary” process, first introduced by Freud in his Interpretation of Dreams (1900/1964, chap. 6). In Freud’s original formulation, secondary process is more or less clearly understood as ordinary wakeful thinking, which is voluntary, discursive, more or less logical, and reality-oriented—a product, in short, of the conscious ego. Primary process, by contrast, Freud regarded as instinct-dominated, oriented only toward gratification of infantile wishes, and rooted in the unconscious id. In Freud’s theory of dream formation, as expounded by Fliess (1959), primary process expresses itself in the various properties of dreams which allow them to serve as acceptable forms of manifest expression for clearly formulated but unacceptable wishes latent in the unconscious. The properties in question include, for example, picturization, symbolization, condensation, concretization, allusion, displacement, and representation through opposites.
The psychoanalytic conception of primary process is approximately convergent, descriptively, with creative imagination as conceived by Myers, Coleridge, Brann, and others. Where it originally differed, of course, was in its profound devaluation of imagination, its relentlessly “archeological” approach to non-verbal symbolism, and more generally its close ties to the sometimes bizarre formulations of orthodox Freudian metapsychology. These differences have been substantially eroded, however, by subsequent revisionary trends within psychoanalysis itself, driven at least in part by trenchant empirical critiques of Freud’s dream theory by Hall (1953) and others. Primary process is no longer conceived as rigidly distinct from and opposed to secondary process, or as a mere slave of the Freudian unconscious—that boiling cesspool of primitive sexual and aggressive urges. Rather, it represents a distinctive mode of cognition existing alongside secondary process, complementary to and interactive with it, and potentially available in service of creative activity. Kubie (1958) in particular has emphasized the versatility, brilliance, and speed of this “preconscious” system, and its role as the wellspring of creativity:
In the adult who is not hamstrung by conscious or unconscious fear and guilt, preconscious processes make free use of analogy and allegory, superimposing dissimilar ingredients into new perceptual and conceptual patterns, thus reshuffling experience to achieve that fantastic degree of condensation without which creativity in any field of activity would be impossible. In the preconscious use of imagery and allegory many experiences are condensed into a single hieroglyph, which expresses in one symbol far more than one can say slowly and precisely, word by word, on the fully conscious level. This is why preconscious mentation is the Seven-League Boot of intuitive creative functions. This is how and why preconscious condensations are used in poetry, humor, the dream, and the symptom, (pp. 34–35)
Even a few mainstream psychologists from Hilgard (1962), Neisser (1963), and Shepard (1978) to Eysenck (1995) and Sloman (1996) have recognized that this psychoanalytic concept of primary process, liberated from its Freudian origins, corresponds to something which is vital in human cognitive makeup generally, and especially significant in relation to creative thinking.21 The same recognition lies at the core of the account by Koestler (1964) of creative inspiration as “bisociation,” or association of ideas across normally separate domains. Indeed, Koestler states flatly that “we find all the bisociative patterns that I have discussed prominently displayed in the dream” (p. 179).
These trends within neo-Freudian psychoanalytic thinking in regard to primary process and creativity have been documented in many places, including Erdelyi (1985), Fromm (1978–1979), Kris (1952), Kubie (1958), Rapaport (1951), and Suler (1980). They coalesce around the concept of “regression in service of the ego,” introduced by Ernst Kris (1952). Stripped of its heavy encrustation in the technical vocabulary of psychoanalysis, this amounts essentially to voluntary, controlled exposure to the free play of imagination (primary process) in service of adaptive purposes such as creativity, wit, and humor (Schafer, 1958). Clearly, this notion is substantially congruent with Myers’s general conception of genius as effecting greater than normal cooperation between supraliminal and subliminal forms of mentation. Kris’s (1952) conception of inspiration likewise is closely similar to Myers’s, except that his clinical perspective and experience compel him to point out the associated psychological hazards as well: “Inspiration—the ‘divine release from the ordinary ways of man,’ a state of ‘creative madness’ (Plato), in which the ego controls the primary process and puts it into its service—need be contrasted with the opposite, the psychotic condition, in which the ego is overwhelmed by the primary process” (p. 60). Koestler (1964, p. 659) speaks in very similar terms of genius as reaching deep underground sources of inspiration through a reculer pour mieux sauter, as contrasted with pathological conditions representing a reculer sans sauter.
The psychoanalytic concepts of primary process and regression in service of the ego provide useful ways of talking about subliminal contributions to creative activity, but they remain primarily descriptive and rely upon a psychological theory which derives mainly from clinical practice and observation. E. F. Kelly (1962) attempted to account for the same descriptive properties in more mechanistic fashion, utilizing Heinz Werner’s (1957) broader-based comparative psychology of mental development. Creative inspiration was again portrayed in much the same terms, emphasizing the emergence into consciousness of pregnant figurative constructs—images, symbols, metaphors, analogies, and the like—organized at an unconscious level; the main novelty, only partly successful, consisted in an attempt to show how such constructs might form spontaneously in accordance with Werner’s account of the mechanisms involved in “primitive” perceptual and conceptual processes.
Common to all these pre-CTM 20th-century accounts was a recognition of the fundamental role of analogizing in creative activity, and more generally of metaphor, broadly construed, as the language of the imagination. Such recognition was hardly original with them, of course: According to Aristotle, for example, the “greatest thing by far is to have a command of metaphor. This alone cannot be imparted by another; it is the mark of genius, for to make good metaphors implies an eye for resemblances” (Butcher, 1951, p. 87). For Emerson as for Myers, analogy represented the fundamental movement of thought, its capacity to penetrate to hidden affinities linking things behind the diversity of their surface appearances. Emerson’s godson William James (1890b) accorded analogy a similarly prominent position within his psychology, declaring the grasping of sameness to be “the very keel and backbone of our thinking” (vol. 1, p. 459). Indeed, for him it lies at the very heart of genius: “Some people are far more sensitive to resemblances, and far more ready to point out wherein they consist, than others are. They are the wits, the poets, the inventors, the scientific men, the practical geniuses. A native talent for perceiving analogies is reckoned by Prof. Bain, and by others before and after him, as the leading fact in genius of every order” (vol. 1, pp. 529–530). Large-scale analogies have unquestionably played an important historical role in shaping scientific thought, two prominent examples being Laplace’s clockwork universe and contemporary efforts to construe the mind as some sort of computational system. At a lower level Draaisma (2000) has documented the role of technological metaphors in the evolution of memory theory, and ironically enough even Perkins (2000), who has explicitly argued against the importance of analogy in creative thinking, develops his “nothing-special” theory in the context of an elaborate and sustained analogy between creative effort and prospecting for gold in Alaska (his “Klondike space” metaphor). The capacity for metaphor, in sum, plays an important role even in scientific thinking, and it represents an important element of our overall cognitive organization.
As noted in Chapter 1, mainstream cognitive science of the last few decades has rediscovered these truths. In addition to the important descriptive studies of metaphor by linguist George Lakoff and his colleagues (Lakoff, 1993, 1995; Lakoff & Turner, 1989), a large amount of computer modeling of analogy processes has been carried out under the auspices of the CTM, work collectively characterized by practitioners Holyoak and Thagard (1995) as having produced “great progress” and constituting what is “undoubtedly one of the success stories of cognitive science” (p. 251).
We respectfully demur, and must now attempt at least briefly to explain why. Our goal here will be to cut through a large and complex literature to what we perceive as its core problem, which is essentially the inability of all existing computational models to address the fundamental issues of semantics or meaning and the intentional activity of knowing human subjects—the heart of the mind (Chapter 1). Non-specialist readers are forewarned that this section may prove hard going on a first reading.
A comprehensive state-of-the-art survey of recent analogy work is provided by Gentner, Holyoak, and Kokinov (2001), but we will begin with two leading exemplars, the structure-mapping theory of Gentner (1983), implemented primarily in the form of their “structure-mapping engine” (SME; Falkenhainer, Forbus, & Gentner, 1989; Forbus, Gentner, & Law, 1994), and the “multiconstraint” theory of Holyoak and Thagard (1995), implemented primarily in the form of their “analogical constraint mapping engine”(ACME; Holyoak & Thagard, 1989) and programs for “analogical retrieval by constraint satisfaction” (ARCS; Hummel & Holyoak, 1997; Thagard, Holyoak, Nelson, & Gochfeld, 1990). There are many significant differences among these models, but for our purposes what matters more is their similarities. We will now sketch these common properties, acknowledging that our description is necessarily terse and that interested readers may need to consult the original sources to gain a fuller appreciation of what we are talking about.
Like all cognitive models inspired by the CTM, these models of analogy processing incorporate two principal components: (1) representations of the model’s knowledge in some cognitive domain or domains, and (2) computational processes that operate on these representations to produce the desired behavior. In almost all of the analogy work, as in classical cognitive science generally, “knowledge” is represented in some sort of predicate-calculuslike logical format which identifies the objects talked about, their properties (the predicates that apply to them), and the relations (possibly multiplace) in which they can participate. Expressions incorporating these elements can serve as arguments in higher-order expressions, permitting very complex propositional structures to be built up. Two important things to note here are: first, that English words are routinely used to identify the various properties, functions, and relations that make up the knowledge representation language; and second, that the representations used by the analogy-mapping programs are typically hand-crafted by their human designers.
SME and ACME are concerned only with interpretation of analogies, and not with their production, which is a far harder problem. “Interpretation” amounts here to identifying the best-possible mapping between knowledge in a source domain and knowledge in a target domain, each characterized by a representation of the above-indicated sort. ACME does this (in what seems to us a somewhat more realistic manner, psychologically speaking) by using a connectionist network to settle on the most coherent way of picking out the cross-domain correspondences between elements of the same type—that is, correspondences of objects with objects, predicates with predicates, and relations with relations—given the available knowledge representations. In this way, for example, Socrates’ description of himself as “a midwife of ideas” is said to be correctly interpreted when this constraint-satisfaction mechanism discovers that the expression:
(1) cause ((helps (obj-midwife obj-mother)) (give_birth_to (obj-mother obj-child)))
which expresses the system’s knowledge in the midwifery (source) domain, maps onto:
(2) cause ((helps (Socrates obj-student)) (knows_truth_or_falsity (obj-stu-dent obj-idea)))
in the Socrates (target) domain in such a way that Socrates bears the same relationship to ideas as midwives do to babies (Holyoak & Thagard, 1989, pp. 344–347). Boden (1991, p. 174) gushes over this result, characterizing it as a step toward understanding what takes place in the mind of a human reader of Shakespeare’s famous passage in Macbeth about “sleep that knits up the ravelled sleave of care,” and Holyoak and Thagard (1995, p. 258) thank her for her “fine overview” of their work on analogy.
We feel obliged to register our astonishment that anyone could take this example, representative of many similar examples in the broader analogy literature, so seriously. To us it seems utter caricature of real human understanding, “explaining” nothing. We feel certain, too, that John Searle would agree with this judgment, finding clear parallels to the hyperbole about computer “story understanding” that originally provoked his critique of the CTM, all of which applies here (Chapter 1).
In this case, however, we can also call upon one further witness who is even more important, because he differs from both Searle and ourselves in sharing the basic philosophical commitments of the CTM, and has himself worked extensively on the subject of analogy. We refer to computer scientist Douglas Hofstadter, who with his “Fluid Analogies Research Group” (FARG) has tenaciously pursued a radically different and highly innovative approach to analogy that is rooted in the study of “microdomains” such as letter-string problems rather than “macrodomain” or high-level analogies of the Socrates/midwife sort. Hofstadter and FARG (1995) provide not only an illuminating history of their own efforts, packed with important psychological and computational insights, but trenchant critiques of their “real-world” competitors, including in particular SME and ACME, as well as many other leading projects in cognitive psychology and artificial intelligence including scientific-discovery and story-writing programs.
We will mention here just a few of Hofstadter’s most telling criticisms of SME and ACME, which are also the ones most important for our own purposes. First, these “real-world” approaches essentially bypass the crucial issue as to how concepts or representations are acquired or constructed in the first place, leaving it to the designers to provide all the necessary “knowledge” in precisely the right form. Both systems lack the dynamic flexibility of human cognition and depend too strongly on the detailed structure of the representations provided them in advance. Hofstadter wonders, appropriately, what would happen if these “source” and “target” knowledge representations were coded independently by different persons. Most importantly, and like many other AI projects, SME and ACME engender a strong “Eliza” effect (Weizenbaum, 1976)—that is, an atmosphere of meaningfulness which depends strongly upon the use of English-like words and expressions in their representational notations. Such notations, even if intended simply as mnemonic aids, covertly engage the semantic capabilities of designers or observers of the system and encourage them to project these capabilities into the system itself (the homunculus problem; Chapter 1). But, Hofstadter argues, SME and ACME are in reality “hollow,” semantically empty; they know nothing about Socrates, midwifery, or anything else, and operate entirely syntactically, in terms of the forms employed in the notation. To underscore this last point Hofstadter suggests replacing the English-like words and expressions of the original notation with letters or numbers. In that case everything would work exactly as before, except that the specious atmosphere of meaningfulness would be dispelled.22
Hofstadter’s critique of these high-level analogy programs, in our opinion, is fair, thorough, and devastating. We wish only to add one further point related specifically to ARCS, the program for retrieving suitable analogs of a given word from long-term memory. Both Boden (1991, p. 174) and Holyoak and Thagard (1995, p. 252) make much of the fact that ARCS utilizes information from WordNet, an electronic thesaurus of English developed independently at Princeton by George Miller and colleagues (Miller, Beckwith, Fellbaum, Gross, & Miller, 1990). Boden in particular suggests that ARCS thereby gains access to a significant part of the conceptual system underlying English, but this suggestion needs to be carefully qualified. As Miller and Fellbaum (1991, pp. 200–201) made clear, WordNet is not “constructive” but merely “differential”‘. It explicitly labels a variety of semantic properties of entries and semantic relations that hold between entries, but it does not attempt to represent the meanings of entries directly; rather, it simply provides brief definitions (“glosses”) sufficient to allow fluent speakers of English to identify meanings which they are presumed already to possess. Although WordNet helps ARCS retrieve words semantically related to a target word, it does so syntactically, by virtue of the labels provided, and has no understanding of word meanings themselves.
If meaning cannot successfully be captured by standard forms of high-level symbolic cognitive architecture, how then are we to deal with it? Hofstadter believes that he has the answer, and that his approach, radically different, ratifies the fundamental connectionist faith that “human cognitive phenomena are emergent statistical effects of a large number of small, local, and distributed subcognitive effects with no global executive” (Hofstadter & FARG, 1995, p. 291). Abandoning the forlorn hopes of those who attempt to get at high-level semantics directly, Hofstadter turns instead to a variety of microdomains in which, he believes, all the essential features of human cognition are present, but in more tractable form. His goal, as the book’s subtitle indicates, is to develop “computer models of the fundamental mechanisms of thought.”
The centerpiece of this work is Copycat, a computer program which generates solutions to analogy problems of the following sort: Suppose the letter-string aabc were changed to aabd; how would you change the letter-string ijkk in the “same way”l No brief description can possibly do justice to the marvelous ingenuity of this program, but we must attempt to convey enough of its flavor to get to our central point (full details can be found in Hofstadter & FARG, 1995; and M. Mitchell, 1993). The program’s architecture includes three main components: First is the “slipnet,” which represents Copycat’s pre-supplied or “Platonic” knowledge and plays a crucial role in its dynamic behavior. This knowledge includes the letters of the alphabet and the numbers 1–5, with each successive pair of each group linked by labeled connections indicating their successor/predecessor relationships, and a variety of other things including pairs such as first and last, leftmost and rightmost, and predecessor and successor, all of which are again explicitly labeled as “opposites” (M. Mitchell, 1993, p. 47). Second is the “workspace,” essentially a short-term or working memory in which candidate solution structures are dynamically assembled in parallel fashion. Finally, the “coderack” contains a large repertoire of “codelets,” essentially micro-level programs or “agents” that launch themselves opportunistically, when their conditions of execution are satisfied by structures appearing in the workspace, or occasionally if they are selected by a random process.23
Each of Copycat’s problem runs is dominated at first by bottom-up or “open-minded” influences consisting of competitive interactions among a host of codelets operating in parallel, but as processing continues and larger candidate solution structures begin to form, top-down or “closed-minded” influences come increasingly into play, reflecting “pressures” created both by the emerging structures themselves and by the changing configuration of the slipnet, which is “rubbery” and dynamically adjusts the “distances” among its concepts in response to the situation at hand. These dynamic properties represent Hofstadter’s way of modeling the context-dependent conceptual halos and “slippages” that he regards as crucial properties of human thinking. A measure of “temperature” falls as the level of “goodness” or order in the workspace increases, but that order itself emerges spontaneously from the collective activity of these low-level processes. There is no central executive.
That will have to do by way of description of Copycat. No one who studies this program can fail to admire its ingenuity, but what is its real significance? Commentators from the “real-world” faction have been inclined to dismiss Hofstadter’s microdomain approach simply on grounds that it seems too tiny, too remote from the high-level problems of interest to them, but Hofstadter defends himself repeatedly and ably against such facile responses. The real core issue, he argues, has nothing to do with the size or subject matter per se of problem domains, but rather with their ability to engender real progress on fundamental issues of cognitive theory, and especially progress on the core problem of meaning.
We agree with this view, and here we come to the crux of the matter: Hofstadter believes there is a fundamental difference between Copycat and the semantically empty “real-world” analogy programs, in that symbols in Copycat, such as “successor,” “acquire at least some degree of genuine meaning, thanks to their correlation with actual phenomena, even if those phenomena take place in a tiny and artificial world” (Hofstadter & FARG, 1995, p. 290). Unlike SME and ACME, he argues, Copycat can survive the label-substitution test—for example, replacing “successor” with “SIGMA”—for the following reason: “An astute human watching the performance of Copycat and seeing the term ‘SIGMA’ evoked over and over again by the presence of successor relationships and successorship groups in many diverse problems would be likely to make the connection after a while, and then might say, ‘Oh, I get it—it appears that the Lisp atom ‘SIGMA’ stands for the idea of successorship’“ (p. 290).
But this core argument, we submit, is specious: In his zeal to sustain the CTM, Hofstadter has simply confused what a conscious, intelligent human observer can learn about Copycat by observing its behavior with what Copycat itself “knows.” He has failed to recognize that the really essential feature of this microdomain consists in the fact that he is able to limit all relevant “meanings” to those which can be transformed into formal or syntactic properties of letters and letter-strings, as defined by the slipnet. Like SME and ACME, Copycat itself “knows” nothing about successorship or anything else. The claimed contrast with real-world analogy programs is therefore not fundamental, and the problems of meaning and intentionality in knowing subjects remain unsolved.24
In sum, neither the “real-world” analogy programs nor Hofstadter’s microdomain approach have as yet done much to illuminate even analogical thinking of quite pedestrian forms, let alone the more dramatic forms of metaphor and symbolism accompanying inspirations of genius.
But the problems go even deeper, in fact, than we have so far indicated: Analogical thinking cannot properly be thought of as some sort of special case, to be handled by mechanisms isolated from the main parts of our cognitive system. Rather, it is simply a relatively conspicuous expression of a cognitive capacity which pervades even the most mundane uses of everyday language, but which has not yet received the attention it deserves. We refer here to the ubiquitous phenomenon of generality in word meaning, discussed in detail in E. F. Kelly (1975). This is something very different from “slippage,” the primary implementation in Copycat of Hofstadter’s important and general ideas about the role of “fluidity” in human thinking. “Slippage” for him means a discrete shift from one sharply-defined concept to a different but related concept—for example, from “successor” to “predecessor”—whether triggered by “pressures” or by a random decision. Generality, by contrast, involves the deployment of a single concept in semantically appropriate ways to a diversity of novel circumstances. This is distinct on the one hand from polysemy, the possession by a general term of multiple distinct senses, and on the other from vagueness, under which we include not only human-type uncertainty about the applicability of a general term to particular cases, but also failures of discrimination, as in a connectionist network.
The main thrust of the lengthy discussion in E. F. Kelly (1975), which we will not repeat here, is to show that generality of word meaning is far more pervasive than most psychologists and linguists realize. Its pervasiveness and importance only become apparent when one is forced to confront large numbers of everyday usages of common words side-by-side, an experience that few contemporary cognitive scientists have had. Note also that this fundamental property of word meaning is closely related to the manner in which memory-knowledge “transcends” or “overflows” its expressions in specific episodes of remembering, as discussed in Chapter 4. Neither phenomenon is adequately accounted for by current cognitive theories of “representation,” and both have been largely ignored in the vast psychological literature on “concepts” (Komatsu, 1992; Medin & Heit, 1999; E. E. Smith & Medin, 1981). We can only concur with Hofstadter and FARG (1995) that “the question ‘what is a concept?’ could be said to lie at the crux of cognitive science, and yet concepts still lack a firm scientific basis” (p. 294).
Meanwhile, what must be fully appreciated is that the domain of metaphor, which scholars have tended to identify with lofty flights of poetic and artistic imagination and to think of as something special, extends in reality downward to the most pedestrian figures of speech of everyday parlance, where it shades over into ordinary generality. As Emerson (1837/1983) remarked, “language is fossil poetry” (p. 457). But there is an evident inner unity linking highest to lowest, and all are expressions in varying degrees of a fundamental and still-unexplained activity of our cognitive organization.25
This already has serious implications for cognitive science, but there is still more to be said. Let us continue for a moment with metaphor and related modes of symbolism. Above all else, these represent attempts to understand, construe, or express something in terms of something else. Philosopher Max Black (1962, 1990) has done a particularly good job, we think, of characterizing this process in a psychologically realistic way. Metaphor breaks rules, but in service of a deeper fidelity. It is inherently creative, for “there can be no rules for ‘creatively’ violating rules” (Black, 1990, p. 55). Good metaphors illuminate; they express and promote insight, through a sort of tension or “interanimation” between individual concepts or large conceptual domains which are kept simultaneously in view. As emphasized particularly by Black, this interanimation often goes far beyond the sort of abstract or literal point-by-point comparison between prespecified and static attributes and relations contemplated by structure-matching theorists such as Gentner (1983). Indeed, and as Myers also suggested, in the most important cases the subject being construed or expressed may itself only become known, or at least better known, as a result of this symbolic process. Bowra (1955) nicely captures this in his continuing description of creative inspiration: “What begins by being almost unconscious becomes conscious; what is at the start an outburst of energy infused with a vague idea or an undifferentiated vision becomes concrete and definite; what is outside the poet’s control is gradually made to submit to his will and judgment” (pp. 5–6).
The critical point here is once again that these more extreme forms of “interanimation” have pedestrian analogs at the core of everyday cognition. This is in fact implied by the phenomenon of generality of word meaning discussed above: Examination of the entries given by desk dictionaries for common words such as “make,” “take,” “turn,” “hand,” or “go” reveals that the number of psychologically real meanings is far lower than the number of usages listed, and indeed those numbers vary enormously among commercial dictionaries for every such word. But if this is the case, comprehension of words in context becomes not a matter of selecting among large numbers of pre-existing, highly specific, and static meanings, but of mutual adjustment or interanimation involving smaller numbers of more general and elastic ones. E. F. Kelly (1975) argued that an appropriate model or prototype for contextual understanding is therefore not enforcement of “selection restrictions” (J. J. Katz & Fodor, 1964), but the philosophical notion of “syncategorematicity,” originally developed in relation to evaluative concepts such as “good.” The basic idea is that the meaning of “good” changes systematically when predicated of violinists, horses, ideas, and so on, but in a manner that is intelligible in terms of the varying natures of the things so labeled. Similarly, the phrases “eat bread” and “eat soup” do not involve different meanings of “eat” (Weinreich, 1966, p. 411) but an adjustment of one more general sense to the different things eaten; a “tiny” elephant is far larger than a “gigantic” mouse, and so on, without end.
Here again we connect with frontline issues in cognitive theory. Whereas most of the work to date on concepts deals only with categorization, the principles by which diverse things are grouped together or recognized as instances of a single concept, we wish instead to emphasize the closely related and even more fundamental problem of conceptual combination. A similar basic impulse clearly underlies other recent developments such as the work of Lakoff and his colleagues on metaphor, the “generative lexicon” of Pustejovsky (1995), the “cognitive grammar” of Langacker (1990, 1999), the “construction-integration” model of discourse comprehension by Kintsch (1998), and the mutual-constraint “coherence” model of Holyoak and Thagard (1995), subsequently generalized by Thagard (1997, 2000) and applied not only to creativity but to a host of related topics in psychology and philosophy. We welcome all these developments as headed in the right direction, but find all of them fatally flawed by their dependence on the grossly inadequate notational devices presently available for representation and combination of “knowledge.” One good example here is that of the real-world analogy programs, discussed above. More generally, most of the work carried out so far on “conceptual combination” hardly gets beyond hackneyed specimens of the “colorful ball” or “pet fish” type. Sooner or later we are going to have to come to grips with the full-blown phenomenology of creative imagination—its coadunative, alembic, and esemplastic powers as conceived by persons such as Coleridge and Myers.
There is one last and even more basic point we wish to highlight here, without much argument, before concluding this section. Cognitive science is also going to have to give up its reliance on the fundamentally flawed idea that semantics can be treated as intrinsic to the sorts of representational structures it employs. In the first instance it is seldom if ever the case that a concept can plausibly be regarded as having a single, fixed structure independent of the circumstances in which it is used. Even more importantly, no such structure, however complex, can fully determine the conditions of its own application. That gap can only be bridged by the intentional activity of a conscious mind with the capacity to use representational forms of all sorts—be they words, phrases, sentences, analogies, metaphors, parables, allegories, images, symbols, charts, maps, equations, diagrams, blueprints, models, sketches in sand, or whatever—to achieve and express insight about states of affairs in the world. These problems are immensely difficult, and despite the efforts of the early Gestalt psychologists (Mayer, 1995, 1999) present-day cognitive science remains very far from solving them. We can again only concur with Hofstadter when he says candidly that “though few seem to recognize or admit it, ours is a field still searching for its foundations” (Hofstadter & FARG, 1995, p. 376). For related discussion see Chapters 1, 4, and 9.
Let us now try to summarize this long discussion of “incommensurability,” all of it precipitated by Myers’s trenchant comments on differences between supraliminal and subliminal forms of mentation. Recent decades have witnessed a welcome renewal of interest among mainstream psychologists in topics such as imagery, analogy, and metaphor that are deeply intertwined with the psychology of genius. In regard to analogy in particular, serious attempts have been made to account for routine “real-world” forms of the phenomenon using conventional cognitive formalisms, but despite claims to the contrary these attempts have produced little if any real progress. The unconventional “microdomain” approach of Hofstadter and FARG (1995) has fared little better, although it is in many ways more interesting. Meanwhile, it has also become evident that the aspects or properties of cognition that make these mid-level problems so intractable in computationalist terms reach not only upward, to the even more difficult phenomena associated with genius, but downward, to the most mundane forms of everyday cognition. In consonance with Wind’s principle, an adequate cognitive psychology of genius will surely accommodate the psychology of everyday cognition, but a truly adequate psychology of everyday cognition is probably not attainable without it.
In this light, more intensive investigation of the cognitive character of subliminal contributions to genius-level creativity is certainly needed. Older psychological theories of creative inspiration, such as those of the various psychoanalysts, as well as E. F. Kelly (1962) and Koestler (1964), have grappled more directly with these core phenomena, descriptively, but fall short of explaining them in a satisfying way. All, significantly, share one central theoretical commitment, which is to seek explanations of creative inspiration entirely in terms of regression to developmentally prior or more “primitive” forms of symbolic activity that persist alongside the more advanced, adult, reality-oriented forms associated with ordinary language. Myers has a different and more radical idea: For him, that “something of strangeness” and the sheer complexity as well as the prodigious memory and speed of mental operations displayed by calculating boys, by automatists such as Hélène Smith and Patience Worth, and by major inspirations of genius, all point to something which transcends ordinary forms of cognition rather than simply preceding them developmentally. We will return to this theme shortly.
Any theory of the creative process entails consequences for a theory of creative personality, and Myers’s unusually rich conception of inspiration as subliminal uprush entails a correspondingly rich variety of implications regarding personality structure, both in geniuses themselves and in humanity at large. His views in this area again not only correctly anticipate the main trends in subsequent research, but in some cases point significantly beyond them.
We can conveniently begin to delineate these implications by reverting briefly to Myers’s discussion of the 19th-century “degeneracy” theorists such as Lombroso, Nisbet, and Nordau. A tradition associating genius with “madness” had of course existed since classical or even pre-classical times, receiving its most influential early expression in Plato’s Phaedrus. The 17th-century poet John Dryden articulated a widely held view with his famous epigram that “Great wits are sure to madness near allied / And thin partitions do their bounds divide.” The degeneracy theorists, however, took this much further. Self-consciously wrapping themselves in the mantle of science, authors of this stripe issued numerous and widely circulated polemics which in effect identified genius with “madness,” “insanity,” or “degeneracy,” claiming that such an identification was fully justified by advances in 19th-century biological psychology.
Myers was one of the few who disputed this facile reductionism, and his reasons were cogent. Beginning in section 315 of HP, he comments directly on Lombroso as representative of the degeneracy movement. There is first a sampling problem, in that Lombroso included along with real geniuses many persons whose eminence derived from sheer hard work or historical accidents rather than the characteristic psychological qualities of genius itself. Second, his piling up of “anecdotes of the follies and frailties” (HP, vol. 1, p. 91) of the eminent is certainly insufficient to establish his central thesis: For one thing, the availability of such stories may reflect little more than the relatively intense scrutiny such persons receive. Subjected to similar scrutiny, Myers points out, “hardly a good easy man among us but might be analysed into half neuropath and half Philistine if it would serve a theory” (HP, vol. 1, p. 91).26
Despite these defects in Lombroso’s procedures, however, Myers recognizes that “there are underlying facts of great importance which give to his view such plausibility as it possesses” (HP, vol. 1, p. 92). What Myers denies is not the existence of a relationship between genius and madness, but its interpretation by the degeneracy theorists. Abnormality in a statistical sense is not necessarily pathological, Myers contends. What matters is the causal structure underlying the observed correlation. For him that correlation reflects the fact that genius and madness share, as an essential common feature, an unusual openness to the subliminal. The degeneracy theorists, however, had missed a crucial difference—namely, that genius masters its subliminal uprushes, whereas the insane are overwhelmed by theirs.27 Genius is not degenerate but “progenerate,” reflecting increased strength and concentration of inward unifying control (HP, vol. 1, p. 70) and increased utilization of subliminal forms of mentation in service of supraliminal purpose. Indeed, in its highest developments genius represents the truest standard of excellence, and a more appropriate criterion of “normality” than conformity to a statistical average.
William James was largely in agreement with these views and expanded on them in the culminating lecture of his Lowell series on “exceptional mental states” (E. Taylor, 1984). Subject to depressive episodes himself, James took special pains to examine in a more detailed way the arguments of the degeneracy theorists, and he identified at least two additional problems of a general sort. First was the issue of overdiagnosis, based on inadequate diagnostic criteria: Against any presumption that hallucinations are invariably diagnostic for psychopathology, for example, James pointed to the demonstrations by Myers and his colleagues that waking hallucinations are not uncommon in the lives of otherwise normal persons (Bentall, 2000; Gurney et al., 1886). Secondly, there are many examples of geniuses, even poetic geniuses, who were paragons of balance and stability. Here James cites Schiller, Browning, and George Sand, along with various members of his own community such as Emerson, Longfellow, Lowell, Whittier, and Oliver Wendell Holmes. Madness therefore is certainly not necessary for genius, although it may sometimes combine with intellect and will to produce it. What madness provides in such cases—what really is necessary, and can also occur in its absence—is uprushes from the “seething cauldron” of the subliminal, as described by Myers.
This basic Myers/James interpretation of relations between genius and mental illness has been confirmed and refined by subsequent research, including not only clinical and biographical investigations but also experimental and quantitative studies of various types. We will next briefly survey some highlights of this large literature.
The clinical side of things was of course at first mainly in the custody of the Freudian school of psychodynamic theorists, which to a considerable extent followed the degeneracy theorists in emphasizing regressive, unconsciously determined, and seemingly pathological aspects of genius to the exclusion of those aspects of rational, conscious control that Myers and James recognized as equally fundamental to success of the creative enterprise. Freud himself was profoundly ambivalent toward genius and the imagination, greatly admiring them but deeply suspicious of their origins (Storr, 1972). Creative imagination, like the dream, was viewed by Freud as substitute gratification, the disguised expression of unacceptable wishes and impulses. Behind it all there is nothing but a deep inability to confront “reality.” Leonardo da Vinci, for example, was characterized by Freud as a repressed homosexual who sublimated his unacceptable urges into scientific curiosity and the production of great works of art. Of course there is little if any historical evidence that Leonardo himself was a closet homosexual, and there have certainly been a great many repressed homosexuals, but few Leonardos. Freud was apparently undeterred by such considerations, however, and legions of his disciples followed his example, producing speculative accounts of wildly varying plausibility of the neurotic origins of various great works of art and literature (see, e.g., W. Phillips, 1957). These psychoanalytic accounts mainly addressed supposed neurotic conflicts of creative geniuses in relation to the thematic content of their works, for as Freud himself had acknowledged in his study of Leonardo, the creative act itself remained unexplained.
We have already described in relation to the creative process how orthodox Freudian theory gradually gave way to reformists such as Kris (1952) and Kubie (1958), who adopted views much closer to those of Myers and James. Kubie’s conception of the “preconscious” in particular has a great deal in common with Myers’s conception of the subliminal; it is the home of imagination, and the vital source of creativity and dreams, rich in affectively loaded symbolism, analogy, and imagery. While retaining Freud’s conception of a truly inaccessible dynamic unconscious which harbors actively repressed and unacceptable wishes and urges, Kubie (1958) treats this as the source not of creativity itself, as in Freud’s view, but of limitations and distortions in an otherwise healthy psychological process: In a nutshell, “neurosis corrupts, mars, distorts, and blocks creativeness in every field” (p. 142). Correspondingly, the creative person is one who enjoys an unusual degree of conflict-free intimacy with preconscious processes, and who thus for Kubie as for Myers is a model of psychological health. Closely similar views in regard to art and neurosis were expressed from the point of view of literary theory in an excellent essay by Lionel Trilling (1953).
The psychoanalytic concept of “neurosis” of course corresponds to relatively mild forms of “insanity” or “madness” as understood by the 19th-century degeneracy theorists. As psychiatry advanced and its diagnostic categories and procedures evolved, attention also focused on relations between genius and more severe forms of mental illness, including in particular schizophrenia and the affective spectrum disorders (manic-depressive or bipolar disorder, major depression, etc.). A sizeable literature developed in which increasingly refined diagnostic procedures were applied either retrospectively, to the biographies or autobiographies of outstandingly creative historical personages, or in clinical-type research with living persons of varying degrees of eminence.
We can quickly summarize the main results of this research, which are clear and consistent (Eysenck, 1995; Jamison, 1993; Juda, 1949; Ludwig, 1995; Nettle, 2001; Post, 1994; R. L. Richards, 1981): There is definitely an elevated incidence of significant mental illness among highly creative persons, and in particular an elevated incidence of affective spectrum disorders among poets, writers, and artists. However, this statistical linkage provides no warrant for any direct or thoroughgoing identification between such illnesses and the psychological conditions of creative activity. Even among poets and writers, for example, where the incidence of affective disorders is highest, it falls far short of universality. Conversely, the vast majority of victims of bipolar disorder and severe depression are certainly not creative geniuses. There may also be some overestimation of incidences, since even modern diagnostic manuals have included aspects of creativity as criteria for the relevant diagnostic entities. The DSM-III used by Jamison (1993), for example, included among its criteria for hypomania “sharpened and unusually creative thinking” and “increased productivity, often with unusual and self-imposed working hours” (p. 265). Furthermore, even when overt psychopathology is incontrovertibly present—as it clearly sometimes is—its causal relation to the creativity is not always clear. The higher incidence of depression among artists and writers, for example, may partly be a consequence rather than a cause of the creative life, owing to the relatively greater risks and uncertainties of such life for them as compared, say, with scientists. Conversely, the suffering occasioned by mental illness may translate more directly for them than for scientists into experiences potentially useful for their creative purposes. There is also the important issue of temporal relations between creative production and outbursts of disorder. Severe and chronic disorders such as schizophrenia, for example, typically put an end to whatever of creativity was present beforehand, and severe depression commonly does so as well, at least temporarily.
Most importantly, a central recurring theme amidst all the reported psychopathology is its coexistence with unusual levels of ego strength, manifested in characteristics such as industry, drive, perseverance, organization, discipline, independence of judgment, tolerance of ambiguity and frustration, and determination to master the subject at hand using any available means of expression (Barron, 1958, 1963, 1968, 1969; Delias & Gaier, 1970, p. 63; Eysenck, 1995, p. 114). Even Jamison (1993, pp. 97–99), despite her somewhat Lombroso-like tendency to emphasize the pathological, clearly acknowledges this characteristic: “There is a great deal of evidence to suggest that, compared to ‘normal’ individuals, artists, writers, and creative people in general, are both psychologically ‘sicker’—that is, they score higher on a wide variety of measures of psychopathology—and psychologically healthier (for example, they show quite elevated scores on measures of self-confidence and ego strength” (p. 97).28
This paradoxical combination of characteristics is evident even in the extreme case of manic-depressive illness, which has been especially common in poets and is Jamison’s special interest. She cites as emblematic Sea-mus Heaney’s extraordinary description, which Myers would certainly have relished, of the poet Robert Lowell:
[Lowell] had in awesome abundance the poet’s first gift for surrender to those energies of language that heave to the fore matter that will not be otherwise summoned, or that might be otherwise suppressed. Under the ray of his concentration, the molten stuff of the psyche ran hot and unstanched. But its final form was as much beaten as poured, the cooling ingot was assiduously hammered. A fully human and relentless intelligence was at work upon the pleasuring quick of the creative act. He was and will remain a pattern for poets in this amphibiousness, this ability to plunge into the downward reptilian welter of the individual self and yet raise himself with whatever knowledge he gained there out on the hard ledges of the historical present. (Jamison, 1993, p. 99)
Genius occurring in conjunction with manic-depressive illness, in sum, is not a model of genius in general but a special case of the more general model set forth by Myers. It has one key characteristic, however, that makes it especially inviting as a subject for further research. Specifically, the stages of the creative process are in this case temporally segregated to a considerable degree, with the “uprush” of imaginative material—the “touch of fire”—strongly linked to the hypomanic phase of the illness cycle.29 This circumstance cries out for within-subject neurophysiological and imaging studies of brain conditions accompanying the cyclic emergence of imaginative activity, as Jamison herself observes (pp. 112–113). Not much has yet been accomplished, unfortunately, along these lines.
Together with the other affective-spectrum disorders, manic-depressive illness also forms part of an even larger picture of biological conditions conducive to creativity. Specifically, Eysenck (1995), Jamison (1993), Nettle (2001), and R. L. Richards, Kinney, Lunde, Benet, and Merzel (1988) have assembled data showing that a capacity for “strong imagination,” which may express itself either in psychosis or in creativity, is at least partly heritable. Thus for example healthy relatives of mental patients show much higher levels of creative accomplishment than members of the general population (J. L. Karlsson, 1984), first-degree relatives of creative persons show higher incidence of mental illness (Andreasen, 1987), and genius, like manic-depressive illness, tends to distribute along family lines even when the descendants are adopted by other families (Jamison, 1993; McNeill, 1971). Exactly what the relevant genetic conditions are, and how they translate into brain conditions conducive to strong imaginative activity, again remains to be elucidated, although Eysenck in particular thinks of imagination primarily in terms of “remote associations” and gravitates toward Martindale’s account of biological conditions conducive to these. Nettle in effect turns degeneracy theory on its head: Not only does psychosis not explain genius, he suggests, but the social value of strong imagination in the form of genius may itself explain the persistence of psychosis, which partially shares its genetic underpinnings.
To summarize the discussion so far, extensive modern research on relations between genius and mental illness has strongly confirmed Myers’s central conception of genius as successful cooperation between supraliminal and subliminal modes of cognition. Further confirmation derives from the neglected subject of “art in the insane,” which we can discuss only briefly here (see also Grosso, in press). Whereas most of us rely excessively on supraliminal modes of cognition, and fail to reach our full creative potential due to insufficient access to the subliminal, these individuals represent the opposite problem—superabundance of subliminal uprush unchecked by the necessary supraliminal control. As Kris reported (1952), “every mental institution in western civilization has its inmates who cover every scrap of paper, every free place on wall or window sill, with words or shapes; there are some who carve with any instrument at hand in any material to which they can gain access or who model in bread if there is no clay” (p. 153).
These phenomena, which have largely been suppressed by the advent of antipsychotic medications, formerly occurred in something like 1–2% of hospitalized patients, usually schizophrenics (Kris, 1952). The creative activity typically began spontaneously, often in persons with no relevant prior training or experience, and went on at great intensity for periods ranging from weeks to decades. The sheer volume of the material, as well as its symbolic richness and complexity, is sometimes staggering: Adolf Wolfli, for example, a paranoid schizophrenic, worked from 1908 to 1930 on a massive narrative that amounts to a personal mythology, a mixture of authentic personal history and cosmic fantasy, interwoven with prose, poetry, pictorial illustrations, and musical compositions. His corpus ultimately comprised some 25,000 packed pages, along with hundreds of drawings, many of which now hang next to the work of Paul Klee in Switzerland (Spoerri, 1997). Morganthaler (1921/1992) was fascinated by Wolfli’s relentless creative output and observed the inmate at his worktable as he wrote and drew. Wolfli would talk about the content of the work but had no idea how he did what he did. Morganthaler observed that he worked from no previous sketch and without conscious plan, yet the entire ongoing oeuvre seemed guided by a single, compulsive, unifying intelligence (MacGregor, 1989).
These phenomena clearly have more to do with automatism than with conscious craft, and belong with the family of mediumistic creativity discussed above. Wolfli shares with Hélène Smith in particular a profile of extraordinary multi-modal creativity powerfully shaped by persistent automatisms. All this of course is consistent with Myers’s notion of subliminal intelligence. Prinzhorn (1972/1995) viewed the phenomena in very Myers-like terms as outpourings of a distinct and fundamental image-making or world-generating capacity of the human mind, whereas Kris (1952) interpreted them from his orthodox Freudian perspective as unusually clear expressions of the primary-process mechanisms at work in ordinary dream-formation. From either point of view, the art of the insane provides a potentially important window on the workings of creative imagination and deserves greater attention in that light.
The brunt of our effort so far has been devoted to showing that Myers’s central conception of genius as successful cooperation between supraliminal and subliminal forms of mentation has been strongly confirmed by subsequent research, and that this has significant ramifications even for contemporary mainstream cognitive science. This is by no means the whole story, however, for his theory of genius as described so far is embedded within a more comprehensive theoretical framework that pushes the envelope of current mainstream thinking about both genius itself and human personality in general even further (Chapter 2). We will conclude by sketching these even more challenging features of his theory of genius.
For Myers, as we have said, genius is “evolutive,” in contrast to the “dis-solutive” phenomena associated with hysteria and other mental disorders. He was profoundly influenced by Darwin, and strove to remain broadly consistent with Darwin’s evolutionary outlook while accommodating the special facts being brought to light by psychical research. Myers portrays genius as the norm of the future, representing a condition of improved psychic integration. The genius thus stands for him among the vanguard on an evolutionary track which humanity as a whole is pursuing, a track that leads “in the direction of greater complexity in the perceptions which he forms of things without, and of greater concentration in his own will and thought,—in that response to perceptions which he makes from within” (HP, vol. 1, pp. 77–78). This evolution, moreover, consists “not only of gradual self-adaptation to a known environment, but of discovery of an environment, always there, but unknown” (HP, vol. 1, p. 95). In Myers’s “cosmical” theory of evolution, this process of progressive discovery itself reflects progressive mobilization of faculties initially latent in the subliminal, including faculties (such as telepathy) which he believed cannot be regarded as products of ordinary Darwinian (“planetary” or “terrene”) evolution.
Although the main lines of Myers’s argument for this more comprehensive and obviously controversial view of evolution are initially deployed in his chapter on genius, we need not attempt to evaluate them here. Myers himself recognized that the phenomena of genius, while consistent with his enlarged conception of the subliminal, cannot in themselves suffice to establish it: “The ‘inspirations of genius’ which seem to spring full-armed into our ken from the depths of our being must count as a form of subliminal faculty, although no theory of such faculty could be based solely upon mental products so closely interwoven with supraliminal or voluntary thinking” (HP, vol. 1, p. 220). Following Myers’s own lead, therefore, we will defer discussion of his larger evolutionary theory to Chapter 9, when we will have a larger supply of relevant facts at hand, and focus instead more narrowly here on its expressions in the realm of genius.
For Myers, the essential link between genius and madness is that they both reflect, in their differing ways, a “perturbation which masks evolution,” an instability caused by the evolving dynamic interplay between supraliminal and subliminal modes of mentation. Genius, however, effects fuller “co-operation of the submerged with the emergent self” (HP, vol. 1, p. 96), and in this way it expresses a nisus (striving or drive) toward greater psychic integration or wholeness that Myers sees as a fundamental property of human nature. Thus, he says, the waking personality “will endeavor to attain an ever completer control over the resources of the personality, and it will culminate in what we term genius when it has unified the subliminal as far as possible with the supraliminal in its pursuit of deliberate waking ends” (HP, vol. 1, p. 152). Genius, in short, draws upon hidden resources of the Self or Individuality in service of a more flourishing conscious life.
Expressions of such a nisus can be discerned in all the various forms and levels of creative activity described above. It can be mobilized, for example, in the treatment of cases of secondary personality (MPD/DID; see our Chapter 5). Thus, Victor Race eventually began to take on the superior qualities of his secondary personality, and early investigators such as Janet and Binet came to understand more generally that their patients’ multiple selves represented opportunities for re-integration, healing, and enlargement of the waking personality. Similar positions have been reached by two thoughtful contemporary students of MPD, Stephen Braude (1995) and Adam Crabtree (1985), from their distinctive vantage points in philosophy of mind and clinical practice, respectively. Crabtree (1985) specifically recommends that the goal of therapy in such cases should be “to assist the creation of a full-blown personality that embodies mixed elements already existing within the individual in a disorganized way” (p. 225).
Similar integrative tendencies can often be seen at work in the related but more benign domain of mediumistic creativity. The highly functional Hélène Smith, for example, said of herself that “I have never been so clearsighted, so lucid, so capable of judging rapidly on all points, as since I have been developed as a medium” (HP, vol. 2, p. 131). No one disputed these observations, according to Myers, and his colleague Flournoy, who studied her extensively, went even further, declaring that those who have looked into mediumship most deeply “see in it a faculty superior, advantageous, healthy, of which hysteria is a form of degenerescence, a pathological parody, a morbid caricature” (HP, vol. 2, p. 132). Prince (1927/1964) responded in similar fashion to Professor Cory’s characterization of Pearl Curran as “disintegrated,” noting that “to all appearances, and judging by her ability to meet the crises and tests of life, she is splendidly integrated” (p. 462). Indeed, Prince continues, “since somehow remarkable literature has resulted, and according to all testimony her own mentality has been improved and her life made happier and more effective since the arrival of ‘Patience Worth’, it seems a pity that we cannot start an epidemic of disintegration.” We generally agree with these sentiments, and we deplore the fact that the scientific study of trance mediumship, so ably begun by persons of the stature of Myers, James, Flournoy, Prince, and others has been so neglected in more recent times. The psychology and neurophysiology of trance mediumship, and its impact on the development of personality, cries out for further investigation using modern research tools.30
A drive toward integration is also observable among creative persons suffering from various forms of overt psychopathology. Even Wölfli, for example, appeared through his writing and art to gain some degree of mastery over the sexual and aggressive impulses that had led to his hospitalization. Writers such as Kubie (1958) and Trilling (1953) discuss in depth how highly creative persons in effect utilize creative work to overcome the fragmentation of their own personalities by neurotic conflicts, and in so doing may help us to overcome ours. Jamison (1993) and especially Storr (1972) articulate similar views regarding the healing power of art, extending them to more extreme forms of mental illness including in particular manic-depressive illness.
Relatively healthy geniuses display a similar nisus, however. Ghiselin (1952) states that in general “the inventor, whether artist or thinker, creates the structure of his psychic life by means of his words” (p. 13). This theme also pervades the discussion by critic Harold Bloom (2002) of 100 geniuses of literature. Genius in Bloom’s view reflects an “aboriginal” compulsion toward the expansion of human consciousness, and thus at the top of his list stands Shakespeare—”a consciousness shaped by all the consciousnesses that he imagined. He remains, presumably forever, our largest instance of the uses of literature for life, which is the work of augmenting consciousness” (p. 12). Storr (1972) takes as paradigmatic a statement by composer Aaron Copland, who, having first explained that the compulsion to create is at bottom a need for self-expression, continues as follows:
But why is the job never done? Why must one always begin again? The reason for the compulsion to renewed creativity, it seems to me, is that each added work brings with it an element of self-discovery. I must create in order to know myself, and since self-knowledge is a never-ending search, each new work is only a part-answer to the question “Who am I?” and brings with it the need to go on to other and different part-answers, (p. 223)
The polymath Johann Wolfgang von Goethe beautifully exemplifies Myers’s view: “What he wanted was totality; he fought the mutual extraneousness of reason, senses, feeling, and will… he disciplined himself to wholeness, he created himself,” says Nietzsche of Goethe in the Twilight of the Gods. This “resplendent instance of jovial genius” himself confirmed, in a letter to von Humboldt, that he strove incessantly toward a synthesis of the practical and imaginative elements of his nature, a condition in which “consciousness and unconsciousness will interact like warp and weft, an image I am so fond of, uniting the human faculties” (Beddow, 1989, p. 111).
It is significant here that Goethe also epitomized for the Swiss analytical depth-psychologist Carl Jung the process of individuation, “the central concept of my psychology” (Jung, 1961/1965, p. 209). We mentioned in Chapter 5 the influence which Myers apparently exerted on Jung, and this concept of a drive to individuation, essentially equivalent to Myers’s nisus, provides one striking expression of that influence. More generally, it would not be too much of a caricature to say that whereas Freud and his school became preoccupied with the “rubbish heap” side of Myers’s broad conception of the subliminal, Jung tended like Myers himself in the opposite direction, toward a preoccupation with the subliminal as “treasure house.” Jung was much more in the grip of von Hartmann and the extremist wing of the German romantic tradition, however, and his writing has a tendency to dissolve into depths of obscurity which we like many others sometimes find impenetrable. Nevertheless, we feel there is much of value in his work, especially in regard to creativity. Jungians in general adopt a far more positive stance than Freud toward imagination, symbolism, and creativity, emphasizing their role in self-discovery and self-integration (Storr, 1972, chap. 16 & chap. 18). In contrast to Freud, for example, Neumann (1959) treats Leonardo’s genius, and genius in general, as a fundamentally transformative and integrative process of personal growth. The Jungian practice of “active imagination” combines creativity with psychotherapy in a manner also consistent with Myers’s view of genius. Described by Barbara Hannah (1981) as the “most powerful tool in Jungian psychology for achieving wholeness” (p. 2), this is essentially a technique for actively engaging the subliminal; one waits, for example, for emotionally charged imagery to well up from below, and then actively attempts to express its meanings concretely in paint, words, gesture, or vocalization.
Similar views regarding the contributions of imagination to psychotherapy, and more generally of creativity to personal growth or “self-actualization,” have been advanced by persons such as Achterberg (1985), Assagioli (1965/1971), Maslow (1968), Sheikh (1984), and Storr (1972), among many others. This leads inexorably to the potentially important but currently vexed practical topic of creativity “training,” around which a large and lucrative commercial industry has in recent decades sprung up. We agree for the most part with critics such as Beyerstein (1999), R. J. Sternberg and Lubart (1999), and Weisberg (1986) that existing training procedures are not grounded in an adequate psychological theory of creativity, and that empirical evidence of their validity and effectiveness is in extremely short supply. Nevertheless, we also feel with Nickerson (1999) cautious optimism that real progress in this direction may be possible—particularly, we suggest, in the context of Myers’s theoretical and practical investigations of subliminal processes.
Myers himself distinctly foresaw this possibility: “Man is in course of evolution,” he says, and “it may be in his power to hasten his own evolution in ways previously unknown” (HP, vol. 1, p. 23). “What advance can we make in inward mastery?” he wonders; “how far extend our grasp over the whole range of faculty with which we are obscurely endowed?” (HP, vol. 1, p. 70). Although Myers himself does not explicitly address these questions, it follows from his general theory that any procedures which encourage increased but controlled interaction with the subliminal can potentially move us in the desired direction. In addition to “active imagination” and creative work themselves, one thinks naturally in terms of cultivating phenomena such as ordinary dreams, lucid dreaming, and hypnagogia, which most persons can probably do, and perhaps—in those with the requisite susceptibilities—activities such as automatic writing or drawing, crystal-gazing or scrying, trance mediumship, and deep hypnosis.
We should also mention in this connection that there is a very large cross-cultural literature dealing with procedures for controlled production of altered states of consciousness and associated phenomena. The shaman, as a specialist in these “archaic techniques of ecstasy,” represents in effect the creative genius of preliterate society (Eliade, 1964; Nettle, 2001). M. Murphy (1992) has situated this form of genius within an even larger context, drawing upon a vast array of anthropological, psychological, and biomedical data to provide a comprehensive natural history of psychophysical transformative practices. From this larger point of view the shamanistic vocation can be seen as one species of a broader genus of personality-development technologies (Grosso, in press). In addition to utilizing particular techniques of the sorts identified above, such personality-development processes might wisely begin by identifying and building upon whatever creative inclinations each person antecedently displays. The range of such inclinations, Myers suggests, could be wide, for “genius may be recognized in every region of thought and emotion. In each direction a man’s everyday self may be more or less permeable to subliminal impulses” (HP, vol. 1, p. 116). Comic genius or wit, for example, has been recognized by authorities as divergent in outlook as Freud, Kris, Bergson, and Koestler as being closely related to the higher forms of genius in its manner of drawing upon subliminal operations.
Creativity in a given domain might also be enhanced more directly, through exposure to outstanding examples of genius in that domain. We have in mind here the obscure but important topic of the response of receptive human observers to great works of art. That something more is involved than purely intellectual appreciation seems certain. Freud in his characteristic way pathologizes this “something extra.” Art provides substitute gratification, a means of escape from reality, and the artist enables us to derive aesthetic pleasure from his fantasies by formally disguising their forbidden origins in the unconscious (Freud, 1920/1968, pp. 384–385). For Jung, in typical contrast, art provides more than aesthetic pleasure; indeed, to the extent that we can imaginatively involve ourselves in a great work of art we vicariously participate in the transformative, integrative process effected by its creator, and are in some measure transformed and integrated ourselves. Some such “resonance” effect may account, for example, for John Stuart Mill’s famous declaration that he was healed by reading Wordsworth’s poetry, and for the fact that Shakespeare “augments” not just his own consciousness, but ours as well. Thus a major goal of creative education in general, as for Harold Bloom (2002), should be “to activate the genius of appreciation” in students, potentiating the transfer of this transformative power (p. 3). It is surely also relevant here that persons who rate high on objective scales of aesthetic judgment also tend to rate high on measures of creativity (Child, 1973, pp. 65–67).
But there is more to be said, for not all works of art have such transformative effects, or have them to the same degree. As theoreticians of aesthetics from Longinas forward have almost unanimously contended, there seems to be a deep inner connection between the psychological phenomenon of inspiration and the transformative power of art. At the level of persons, for example, it is simply an empirical fact that virtually every major contributor to the western cultural canon who has commented on the creative process has spoken appreciatively of the role of inspiration (Ghiselin, 1952). And the relationship goes even deeper, apparently, for it is typically not their creative work in general but their best work that is described as so benefitting.
That this is not simply the opinion of the creator but also involves qualities inherent in the creative product itself is argued explicitly in the case of poetry by Bowra (1955, pp. 19–25). “Poems which are known to have been conceived by their authors in inspired moments are often those which move us most powerfully,” he says, stating a potentially testable hypothesis. Furthermore, “if we look into the question, we see that inspired poetry has certain qualities which are responsible for its hold on us and for our continued delight in it.” The qualities that Bowra goes on to identify need not occupy us in detail, for they overlap strongly, and unsurprisingly, with the qualities of creative imagination as delineated by persons such as Coleridge and Myers.
And here we return at last to the main thread of Myers’s own explicit argument—the thread we temporarily put aside at the end of our discussion of “incommensurability” between supraliminal and subliminal modes of expression. As we indicated there, Myers does not believe that creative imagination can be fully explained in terms of facts of personal biography, developmentally earlier or more “primitive” forms of symbolic activity, and the like. In the last 14 pages of the genius chapter he develops his alternative view, according to which the roots of imagination, of expression, and indeed of genius itself, lie much deeper.
The essence of Myers’s conception is that both human personality and nature as a whole have complex, multilevel organizations, that these organizations are in some meaningful sense “parallel,” and that “inner” and “outer” somehow mingle at the deepest level. Like Coleridge, who warred incessantly against “the despotism of the eye,” Myers believes that ordinary supraliminal perceptual and cognitive processes reveal only relatively superficial aspects of the far wider and deeper environment, mostly unknown, in which we are continuously immersed. The subliminal reaches further into this complex reality, however, and can report what it finds using its own characteristic modes of symbolic expression. Thus genius, the distinctive characteristic of which is “the large infusion of the subliminal in its mental output” (HP, vol. 1, p. 97), provides means for discovery of this hidden environment.
Views of this sort were of course especially characteristic of the 19th-century English romantic poets (Abrams, 1953; Bowra, 1955; Prescott, 1922), although contemporary American writers of the “transcendentalist” school held similar views. Emerson (1837/1983), for example, says of the genius that “in going down into the secrets of his own mind, he has descended into the secrets of all minds” (p. 64); and not just of minds, but of nature as well, for “nature is the opposite of the soul, answering to it part for part. One is seal, and one is print. Its beauty is the beauty of his own mind. Its laws are the laws of his own mind” (p. 56).
Among later psychologists, the figure who comes closest to Myers is again Carl Jung (1952), who speaks like Myers of true symbolic expression as “the expression of something existent in its own right, but imperfectly known” (p. 215). Jung specifically argues that in the highest or “visionary” mode of poetic creation, epitomized for him by Part 2 of Goethe’s Faust, what is expressed derives from the deepest stratum of the psyche, a transpersonal region which he calls the “collective unconscious,” where inner and outer come together (see also Neumann, 1959). We ourselves think that the sorts of comparative and cross-cultural investigations of “archetypal” symbolism carried out both by Jung himself and by others such as Joseph Campbell are broadly consistent with, and potentially enrich, Myers’s conception of the subliminal and its operation, but it cannot be denied that Jung’s own writings about archetypes and the collective unconscious are especially obscure, and certainly they have had little if any impact on mainstream psychology.
Why then should the mainstream pay more attention to Myers than to Jung in regard to transpersonal aspects of genius? In a nutshell, because here as elsewhere Myers’s system is far better grounded empirically, as demonstrated both by the present book and by HP itself. In his characteristic way, Myers begins in section 331 of HP to marshal arguments for his enlarged conception of the roots of genius. Other parts of HP serve to demonstrate that subliminal products are prone in general to be carriers of supernormal knowledge, and the question therefore naturally arises whether this is also true of the subliminal uprushes associated with genius.
Myers asks first whether such uprushes, like those of a medium or sensitive, tend to display supernormal knowledge of “definite facts.” There is of course a tradition going back to classical times that links these two domains: The Muses, for example, were the daughters of Memory and possessed the gift of prophecy. Although Myers recognizes that these gifts may be somewhat separable in the subliminal region, he clearly expects them to overlap to some degree, and he declares unequivocally that remarkable instances of supernormal knowledge in persons of genius “undoubtedly do exist” (HP, vol. 1, p. 108). He declines to discuss “the most conspicuous and most important” of such instances, by which he means events in the lives of the great personages of various religions, and he refers briefly instead to two other historical geniuses—Socrates and Joan of Arc—whom he himself had investigated extensively. Both cases, details of which are provided elsewhere (Myers, 1889a; HP, vol. 2, pp. 95–103), involved conspicuous episodes of trance and automatism, accompanied (more clearly in the case of Joan) by infusions of definite supernormal knowledge.
Myers drops the subject at this point and moves on, but in light of subsequent developments more can now be said. First, many additional instances have come to light of spontaneous psi experiences in the lives of eminent persons, a particularly good collection being that of Prince (1928/1963). These demonstrate clearly that psi occurs in highly creative persons, but not necessarily that it occurs more frequently or dramatically in them than in the less creative. Two lines of modern experimental evidence, however, more clearly confirm Myers’s expectation of linkages between creativity and psi.
Both trace back to a seminal paper by R. A. White (1964), who sought to identify psychological conditions associated with the unusual levels of success in some older “free-response” ESP experiments (these had used materials such as objects and drawings, rather than ESP cards or the like, as the targets). Her resulting “recipe” for psi success strikingly parallels the Wallas/Hutchinson model of the creative process: Specifically, following an initial period of sensory isolation and relaxation, it prescribes a deliberate demand for the needed information (preparation) followed by waiting and release of effort (incubation), spontaneous emergence of information into consciousness (inspiration, usually in the form of involuntary visual or auditory imagery), and evaluation or elaboration of the information so retrieved (verification). White’s analysis was one of the main influences leading to Honorton’s (1977) model of psi and internal attention states, and to the large body of successful psi research using “Ganzfeld” procedures (Alvarado, 1998; Braud, 1978; see also the Appendix).
A number of additional experimental studies have sought more directly to measure relationships between creativity and psi performance. Both creativity and psi have generally been in short supply in these studies, unfortunately, and most have relied excessively, in our opinion, on psychometrically suspect “creativity” measures of the conventional sorts (“divergent thinking” and the like; see, e.g., Barron & Harrington, 1981; Delias & Gaier, 1970; Hocevar, 1981). Nevertheless, some order has begun to emerge, particularly since the advent of the Ganzfeld procedures. Free-response Ganzfeld studies carried out by a variety of investigators and laboratories have now also clearly indicated that “creative” groups such as musicians, artists, and writers perform conspicuously better than unselected subjects, averaging something over 40% direct hits where only 25% would be expected by chance (Dalton, 1997; N. J. Holt, Delanoy, & Roe, 2004; Schlitz & Honor-ton, 1992).
In sum, although much remains to be done to sort out details of the relationship between creativity and psi, that such a relationship exists now seems beyond reasonable doubt. We emphasize again that this confirms a specific expectation flowing directly from Myers’s theory of genius.31
Beyond occasionally providing supernormal access to definite facts of the ordinary kind, the subliminal uprushes of genius reveal, Myers thinks, that we are “capable of a deeper than sensorial perception, of a direct knowledge of facts of the universe outside the range of any specialised organ” (HP, vol. 1, p. 111). In support of this thesis he appeals first to the poet Wordsworth, and especially to his unique autobiographical poem The Prelude, published posthumously and dedicated to his friend Coleridge (see Wordsworth, Abrams, & Gill, 1979). “Most poets have been Platonists,” Myers observes, and “their influence tends to swell that ancient stream of idealistic thought which lies at the root of all civilised religions” (HP, vol. 1, p. 109). Wordsworth, a member of Bloom’s top 100, is typical in experiencing his moments of inspiration as moments of entrance or insight into a normally hidden spiritual environment that somehow undergirds or interpenetrates the everyday, observable world. By means of the “auxiliar light” of his poetic imagination, Myers suggests, Wordsworth obtained “a vague but genuine consciousness” of this invisible world. His poetry, which had earlier provoked Coleridge’s extensive investigations of the creative imagination, epitomizes for Myers the symbolic process essential to all great art, “which abandons logical definiteness of statement for the sake of a nearer approach to truth hidden in the ideal world” (HP, vol. 1, p. xxx). Wordsworth’s testimony on these matters was especially convincing to Myers, not merely because of its unmatched volume and evident candor, but because “he, if any man, has kept his mind, as Bacon advised, concentric to the universe” (HP, vol. l,p. 111).32
The outreach of the subliminal, moreover, is not purely intellectual or “telaesthetic” in character, for it is motivated and guided by affective factors, and in particular by a sense of beauty. Here Myers appeals to Plato himself, drawing upon his conception of love as expressed by the prophetess Diotima in the Symposium (192–212). Love for Myers is the foundation of both genius and religion, religion being conceived in its essence as genius of the spiritual realm. Love draws the genius ever forward, ascending stage by stage in Plato’s characteristic fashion from beautiful objects to beautiful ideas and ultimately to the archetype of absolute beauty itself, whereupon “beholding beauty with the eye of the mind, he will be enabled to bring forth, not images of beauty, but realities (for he has hold not of an image but of a reality)” (Jowett, 1892/1937, p. 335). For Myers as for Plato, love, beauty, and truth ultimately merge in a state of exalted consciousness.
Having expressed in such soaring terms his deepest personal intuitions, Myers concludes by contrasting his view of genius with the conventional one, reverting for this purpose to his “diagrammatic” example of genius, the calculating boy. What is the ultimate source of this strange ability, he wonders. He acknowledges that some sort of spontaneous genetic variation may be involved; indeed, this hypothesis, “being hardly more than a mere restatement of the facts,… cannot help being true, so far as it goes” (HP, vol. 1, p. 117). But Myers doubts that the biological variation itself can explain those facts. His own answer is that the ultimate source of Dase’s computational faculty lies in some sort of ideal world in which “the multiplication table is, so to speak, in the air” (HP, vol. 1, p. 119). Thus, as in the case of other supernormal phenomena of subliminal origin, biological variation did not create but only revealed Dase’s ability: “By some chance of evolution—some sport—a vent-hole was opened at this one point between the different strata of his being, and a subliminal uprush carried his computa-tive faculty into the open day” (HP, vol. 1, p. 119).33 The progress of evolution, epitomized in genius, consists largely in the progressive revelation and integration of such latent subliminal faculty and, in this way, fuller discovery of this ideal world.
Myers ends here, acknowledging that his grand theoretical vision of genius as personality in evolution goes beyond what the science of his day can guarantee, and returns to the sober exposition of matter-of-fact issues that characterizes the great bulk of his book. How does all this relate, we must now ask, to a contemporary psychology of genius?
It probably could go without saying that few if any mainstream psychologists of the past century would have taken these more challenging elements of Myers’s theory of genius at all seriously, had they ever been exposed to them. Nor have the various humanistic and transpersonal psychologists who have advocated Myers-like views from the margins of the field shown much awareness of the additional strength their positions could derive from Myers’s relentlessly empirical approach to an enlarged scientific psychology. Both groups, we suggest, would do well to revisit Myers’s theoretical challenge in light of more recent developments.
To begin, the Neoplatonic elements in Myers’s account of genius no longer appear as radical as they might have even a few decades ago. The sense of beauty, for example, has increasingly been recognized as playing a vital role in creative activity in all fields from mathematics and science to the arts. Koestler (1964) in particular had urged this view upon the early cognitive psychologists, without much success, declaring for example that “beauty is a function of truth, truth a function of beauty. They can be separated by analysis, but in the lived experience of the creative act—and of its re-creative echo in the beholder—they are as inseparable as thought is inseparable from emotion” (p. 331). A. I. Miller (2001) documents in detail the role played by a sense of beauty in both Picasso and Einstein. Even Poin-care, in what surely is one of the most mechanistic accounts ever provided of the creative process, invoked the notion of a subliminal aesthetic “sieve” that would only pass through to waking consciousness, among the innumerable “combinations of mental atoms” mechanically formed by unconscious processes, those whose “elegance” would make them of real mathematical interest. The contemporary computationalist Hofstadter makes the point even more directly, saying “I feel that responsiveness to beauty, and its close cousin, simplicity, plays a central role in high-level cognition, and I expect that this will gradually come to be more clearly recognized as cognitive science progresses” (Hofstadter & FARG, 1995, p. 318). That it is in fact being so recognized is evident, for example, from R. J. Sternberg and Davidson’s (1995) book on the nature of insight, in the foreword to which Janet Metcalfe welcomes this as something novel; many of the contributors, she says, “remark on the joy that is felt with creative insights and also on the appreciation for beauty that characterizes creative people. This aspect of the phenomenology of insight, until very recently, has been neglected as being too warm and soft for cool, hard Cognitive Science” (p. xi).
Most mainstream psychologists will of course hold out hope that the sense of beauty itself will ultimately prove within reach of conventional biologically-based thinking. We doubt it, but will not argue that point here. Much more serious problems, however, are posed by the “telaesthetic” aspect of creative intuition, its capacity to penetrate beyond the current horizon of knowledge to real “facts of the universe.”
Myers in fact considerably underestimated the strength of his case in this area, for poets are not the only Platonists, and the products of creative intuition are often anything but vague. We refer here especially to pure mathematicians and mathematical physicists, nearly all of whom are Platonists of some description. For Einstein, for example, music and physics were closely linked:
He imagined Mozart plucking melodies out of the air as if they were ever present in the universe, and he thought of himself as working like Mozart, not merely spinning theories but responding to Nature, in tune with the cosmos…. He thought of both musical and physical truths as Platonic forms that the mind must intuit. Great music cannot be “created” any more than great physics can be deduced strictly from experimental data. Some aesthetic sense of the universe is necessary for both. (A. I. Miller, 2001, p. 186)34
Another Nobel laureate in physics, Eugene Wigner (1960), has spoken of “the unreasonable effectiveness of mathematics,” its ability to describe fundamental processes in the physical world to astonishing levels of accuracy. This Platonic theme has been developed most vigorously and explicitly in recent years by mathematician Roger Penrose (1989, 1994, 1997), chiefly in relation to mathematical concepts although he clearly presupposes that such a Platonic world may contain “other absolutes, such as the Good and the Beautiful” as well (Penrose, 1997, p. 1).
Further examples can readily be found in the realm of pure mathematics. Hadamard (1949, chap. 8) briefly discusses some cases of “paradoxical” intuition, including in particular discoveries by Fermat, Riemann, and Galois of correct mathematical results that were not obvious and that went far beyond any possibility of proof by means of the mathematics available at the time. Indeed, some of the proofs were only accomplished following development of whole new areas of mathematics over periods ranging from decades to centuries. Hadamard himself specifically suggests, like Myers, that such intuitions arise from unusually deep strata of the psyche, and that they sometimes emerge in the form of automatisms.
These properties are also found in the extraordinary and well-documented life of the modern mathematical genius Srinivas Ramanujan (Eysenck, 1995, pp. 195–201; Kanigel, 1991; Newman, 1948). There were no recognizable mathematicians in Ramanujan’s family, and he received only a patchy formal training in the course of a generally unhappy educational experience in the schools of his native south India. Nevertheless, between the ages of 16 and 26, ignited by what amounted to little more than a dry compendium of some 5000 known mathematical equations, this largely self-taught prodigy managed not only to recapitulate single-handedly a sizeable fraction of the history of Western mathematics but to generate an astonishing volume and variety of novel results in number theory as well. Discovered in 1913 by the distinguished British mathematician G. H. Hardy, Ramanujan continued this prodigious outpouring until his untimely death in 1920 at the age of 33. Some of his most important theorems have already taken decades to prove, and his crammed notebooks will continue to occupy mathematicians for generations to come. His work has found application in areas as diverse as blast-furnace design, manufacture of plastics and telephone cables, cancer research, statistical mechanics, and computer science. On Hardy’s informal scale of natural mathematical ability, on which most of us would rate close to 0 and Hardy placed himself only at 25, the magnificent David Hilbert ranks an 80, and Ramanujan stands all by himself at 100.
All the main ingredients of Myers’s conception of genius are conspicuously present in this case. First there is extraordinary memory; Hardy recounts, for example, that upon his informing Ramanujan of the number of a taxi in which he had just arrived for a visit, Ramanujan exclaimed at once that this number, 1729, is the smallest integer expressible as the sum of two cubes in two different ways.
Second and more important, his biography is replete with signs of automatism. Some examples: “It was the goddess Namagiri, he would tell his friends, to whom he owed his mathematical gifts. Namagiri would write the equations on his tongue. Namagiri would bestow mathematical insight in his dreams” (Kanigel, 1991, p. 36). “Another time, in a dream, he saw a hand write across a screen made red by flowing blood, tracing out elliptic integrals” (p. 66). This appearance of drops of blood in his dreams signified the presence of the god Narahimsa; then “scrolls containing the most complicated mathematics used to unfold before his eyes” (p. 281). Unfortunately, neither Hardy nor apparently Ramanujan himself took much interest in observing or reporting these psychological phenomena.
Ramanujan’s theorems were “elegant, unexpected, and deep” (p. 206). Mathematicians of great ability, including Hardy among many others, were “enraptured” by his work, and specifically by “its richness, beauty, and mystery—its sheer mathematical loveliness” (pp. 349–350). He was not often wrong, and even when he was wrong (as in some early work on the distribution of prime numbers), the incorrect results still exuded this peculiar atmosphere of mathematical beauty. Yet, as Hardy himself observed, “all his results, new or old, right or wrong, had been arrived at by a process of mingled argument, intuition, and induction, of which he was entirely unable to give any coherent account” (p. 216).
Ramanujan was also an overt Platonist, Indian style. “He pictured equations as products of the mind of God” (p. 282). Mathematical reality exists independently of us, and is discovered, not made; “for him, numbers and their mathematical relationships fairly threw off clues to how the universe fit together. Each new theorem was one more piece of the Infinite unfathomed” (p. 66). In this matter at least, Ramanujan, the apotheosis of intuition, was in complete accord with his colleague and mentor Hardy, the master of proof.
Creative imagination on this scale, we submit, fairly beggars the theoretical apparatus available to contemporary cognitive science, its “associations,” predicate-calculus “representations,” and all the rest. We make no exception here of the recent attempt by Lakoff and Nunez (2000) to account for the origins of mathematics in terms of Lakoff’s more general theory of “embodied cognition.” In brief, these authors argue that a fundamental cognitive process of “conceptual metaphor” can fully explain the progressive elaboration of systems of increasingly abstract mathematical ideas, both historically and ontogenetically. From this point of view, moreover, they summarily dismiss the Platonic “romance of mathematics” as mere pre-scientific mythology, a folk-theory, immune to empirical evidence and acceptable only as a matter of a priori faith.
Apart from its pedagogical implications, which appear substantial, we find this theoretical position deeply flawed, and if writing a full review we would develop the following as our main points of rebuttal: (1) Lakoff and Nunez insufficiently appreciate the fact that inventions of pure mathematics have often been discovered only long after the fact to have profound physical parallels. (2) The realization of a concept such as “number” in alternative and seemingly inconsistent mathematical forms does not invalidate a Platonist view, since even on their own terms this is compatible with the possible existence of the “idea” of number at some deeper psychological or ontological level. (3) More generally, they persistently confound issues related to the ontological status of mathematical ideas with issues as to the means by which we acquire them. (4) They also persistently caricature the Platonist position, claiming for example that it implies that mathematical ideas are literally in or part of the observable world, and that classical symbolic computationalism is the correct theory of the mind; yet the Platonist mathematician Roger Penrose, who explicitly rejects both of these views, is nowhere discussed. (5) On their own side, meanwhile, they consistently overstate the progress of cognitive science generally, and of George Lakoff in particular; simply to label the phenomenon of “conceptual metaphor,” for example, does nothing whatever to explain it. (6) Finally, a plethora of ex cathedra pronouncements scattered through their book demonstrates that Lakoff and Núñez themselves are guilty of precisely the same sort of faith-based a priori theoretical commitment they attribute to their Platonist adversaries: Specifically, they themselves accept, without question, the current mainstream dogma that everything in the mind must be generated by brain processes, period. But this quasi-religious doctrine, which they incorrectly claim “has been scientifically established by means of convergent evidence within cognitive science” (p. 363), is itself falsified by a wider range of empirical evidence of the sorts originally marshaled by Myers and updated in the present volume. Within that larger framework, we conclude, some sort of “Platonic” construal of mathematics remains an empirically viable possibility.35
With this as background, then, let us return briefly to Ramanujan. Even the normally ebullient Hans Eysenck (1995) sheepishly acknowledges that this case “shows just how much distance there is between that which we have to explain, and our puny achievements in explaining it” (p. 201). Kanigel himself, struggling to come to terms with the extreme character of Ramanujan’s genius, begins to wonder whether it is entangled somehow with his pronounced mystical streak, and his general immersion in south Indian mystical culture. He notes the consistency of such a view with longstanding majority tradition in both East and West, and cites approvingly a passage from Hadamard which explicitly links that tradition to both Myers and James:
That unconsciousness may be something not exclusively originating in ourselves and even participating in Divinity seems already to have been admitted by Aristotle. In Leibniz’s opinion, it sets the man in communication with the whole universe, in which nothing could occur without its repercussion in each of us; and something analogous is to be found in Schelling; again, Divinity is invoked by Fichte; etc.
Even more recently, a whole philosophical doctrine has been built on that principle in the first place by Myers, then by William James himself…. According to that doctrine, the unconscious would set man in connection with a world other than the one which is accessible to our senses and with some kinds of spiritual beings. (Hadamard, 1949, pp. 40–41)
What Myers and James really said, of course, is not that the subliminal connects us to another world, but that it connects us to the one existing world in novel ways. The more extreme phenomena of genius, and especially phenomena of the sorts adduced by Myers and amplified here, show that the roots of genius reach into or through a part of our being which somehow transcends ordinary biological functioning in space and time. The deepest affinity of genius lies with this transcendent or transpersonal aspect of our psychological constitution, and hence with the world-wide phenomenon of mysticism.
It is no accident, therefore, that Myers selects as his culminant example of genius, the completest type of humanity, “the eagle soaring over the tomb of Plato”—the Alexandrian mystical philosopher Plotinus (HP, vol. 1, p. 120). But before proceeding further with this theme we must take special pains to correct a misapprehension which is already rampant in the psychological literature. Notwithstanding assertions to the contrary by R. J. Sternberg and Lubart (1999), Boden (1991), Weisberg (1986), and many others, mysticism does not imply supernatural intervention. It is true that in pointing out these psychological connections with mysticism Myers hews close to the classical origins of the terms genius and creation, with their well-known supernatural connotations. But the essence of what he is doing is to respect the impressive phenomenology of genius—reflected in the concept of “inspiration” as being literally “breathed into” by the Muses, a god or daemon, or whatever—while reinterpreting it in entirely naturalistic, functional terms. As we will show in the following chapter, the term “mystical” properly refers to a very large and important class of real human experience which is not necessarily or even primarily “religious.” It does not deserve to be used pejoratively, and it is not synonymous with “unscientific.” To attempt, as many do, to dichotomize accounts of genius into those that are mystical, as opposed to those that are supposedly scientific, is altogether specious.
From a purely historical point of view, Myers correctly anticipated most of what we think has been best in the first century of psychological work on “creativity.” His basic conception of the creative process as successful cooperation between supraliminal and subliminal modes of cognition, for example, anticipated the most mature developments in 20th-century psy-chodynamic theorizing, and led him to an interpretation of relationships between creativity and mental illness that has been strongly confirmed in subsequent research. In a qualified sense his emphasis on “continuity” also anticipates the nascent efforts by present-day cognitive theorists to characterize creativity in terms of extraordinary use of ordinary cognitive abilities.
Still to be taken up, however, are a variety of further challenges which Myers’s theory of genius poses both for the psychology of genius itself and for psychology more generally. All derive from the fact that Myers, unlike most contemporary mainstream psychologists, deliberately focused on genius in its most extreme manifestations rather than on “creativity” of the more readily available forms. The account of genius which he developed in this way has implications in at least four key areas:
The first concerns the need for cognitive psychology to expand its scope to deal more effectively with non-verbal forms of expression in general and with full-fledged creative imagination in particular. Interest in these matters appears to be increasing, but the “imagery debates” and so on of recent decades constitute in our view only the barest beginnings of such an enlargement. Important but neglected resources here include studies of the creative imagination by Coleridge and the romantic poets, investigations of “primary process” thinking by psychodynamic theorists, and investigations of the hallucinatory syndromes induced by psychedelics and various other means.
The second concerns the homunculus problem, and our persisting inability even to recognize, let alone explain, the ubiquitous phenomena of intentionality, seeing-as, and insightful grasping of meanings as central aspects of human mental life. These phenomena occur in conjunction with all forms of symbolic activity, both discursive and non-discursive, from the highest flights of poetic metaphor and symbolism to the altogether mundane phenomenon of generality of word meaning in ordinary discourse. Cognitive science has a serious “inattentional blindness” problem, we believe, and meaning is the unseen gorilla in the room (Simons & Chabris, 1999).
A third challenge involves Myers’s conception of nisus and the transformative power of creation both for genius itself and for those who interact with its products. Although this theme too is already present in the literature to some extent, we think much more research could be carried out along these lines. Myers’s work leads both directly and indirectly, for example, to a variety of novel possibilities for creativity enhancement that merit further investigation. We see this as one aspect of the larger subject of “transformative practices,” which we will address more fully in the following chapter.
Finally, Myers led the way in exposing an inescapable connection between genius and mysticism at the foundations of human personality. Mysticism, even more than genius itself, has been a topic actively avoided and even despised by most mainstream psychologists, but like Myers and James we believe that it will ultimately prove not only significant but central to an empirically adequate scientific psychology. We turn next to this.
1. A striking exception to this widely circulated degeneracy interpretation of genius is the book by German psychiatrist William Hirsch (1896), translated anonymously into English by James’s philosophical colleague C. S. Peirce. That James himself held views similar to and strongly influenced by those of Myers is apparent from his 1896 Lowell Lectures, as reconstructed by E. Taylor (1984).
2. E. Taylor (1984, pp. 162–163) uses this very passage—almost verbatim and without identifying it as Myers’s—by way of characterizing Myers’s theory of genius as the chief background for James’s final Lowell Lecture.
3. Wallas identified the first three stages in an 1891 lecture by Helmholtz, on the occasion of his 70th birthday, in which he described how his best scientific ideas had come to him, and the fourth in the famous account by Poincare of some of his mathematical discoveries (see Hadamard, 1949, or Ghiselin, 1952). Hutchinson arrived at his formulation through study of hundreds of accounts involving not only scientists and mathematicians but artists, musicians, and literateurs, mostly persons of lesser magnitude than Helmholtz and Poincare.
4. Ellenberger (1970, pp. 150, 314–318) specifically credits this term and concept to Myers (although he uses the alternate spelling “mythopoetic”) and laments its general neglect by subsequent workers in dynamic psychiatry. A major exception is Myers’s colleague and admirer the Swiss psychologist Flournoy, whose studies of the medium Hélène Smith we will come to shortly.
5. See also our Chapter 1. Myers states clearly that he would rather have used illustrations drawn from higher mathematics, were sufficient data available, and he appeals to mathematicians for accounts of the mental processes accompanying attainment of their highest results (HP, vol. 1, pp. 78–79). This may have precipitated a subsequent survey of mathematicians carried out by Claparedè and Flournoy, which in turn stimulated the famous account by Poincaré in 1904 of his inventive processes (see Hadamard, 1949, pp. 10–11). Hadamard’s own book answers Myers’s appeal more definitively, combining Poincare’s account with a substantial number of others, and strikingly confirms Myers’s general outlook.
6. Various other kinds of supernormal abilities have also been reported in such individuals. For example, when Sacks accidentally dropped a large box of matches on the floor, both twins instantly perceived that there were 111 matches in all, and that this number is the product of the prime numbers 37 and 3. According to Scripture (1891, pp. 20, 39–40), this unusual capacity for immediate perceptual grasping of numerosity (“subitizing”) is universal among calculating prodigies. Dase, for example, could instantly determine the number of sheep in a herd, or of books in a book-case, or of window-panes on the side of a large house. Yet for most of us such judgments remain accurate only to something at best on the order of the usual 7 ± 2 items (Mandler & Shebo, 1982; G. A. Miller, 1956). Rimland (1978, and personal communication, November, 1978) has also reported finding in his survey of savant-type skills in early infantile autism a few cases in which recurrent spontaneous psi phenomena were conspicuously present.
7. Myers gives a condensed account of this work in his chapter on hypnotism (HP, vol. 1, pp. 194–195, plus supporting appendices), but a much fuller account, including descriptions of the precautions against various potential sources of error, appears in Bramwell (1903, pp. 114–139).
8. There may also be an “absolute” time-sense, analogous to the well-known sense of absolute or perfect pitch which is common in the musically gifted (though neither necessary nor sufficient for high achievement) and apparently universal among musical savants (Treffert, 1989). One of us (EFK), as a teenager, was briefly acquainted with a younger boy, in other respects ordinary, who always somehow “knew,” with startlingly small error, what time it was. Treffert (1989, pp. 97–98) describes two similar cases, one involving “Ellen,” a blind musical savant who also displayed spontaneous ESP abilities.
9. Klüver himself was not entirely clear or consistent about this classification, which seems to us somewhat arbitrary, and Siegel (1977) and colleagues, working mainly with psychoactive drugs, subsequently identified additional form constants as well as a variety of color and movement constants. These details need not concern us here, however.
10. For further details see Sacks (1999) and Wilkinson (2004).
11. Sacks’s own worries remain at the level of Klüver’s form-constants, and he seeks to ameliorate them by appealing to the modern vogue of non-linear dynamic systems, conceived as operating within the visual system itself to produce more global forms of self-organized activity that are not so closely tied to the hard-wired cortical architecture and thus might capture the required properties. This approach barely begins to address the real complexities of the problem, as indicated in the text, and even on its own narrower terms the simulation results provided by Sacks are hardly very encouraging. The results of Bressloff et al. (2001) are far more impressive, and yet these authors explicitly acknowledge various limitations of their model.
12. The success rate of the Ganzfeld might actually prove much higher, we surmise, if attention were confined more narrowly to those subjects and conditions that produce altered states of consciousness with hypnagogic-like, vivid, and autonomous visual imagery, as contrasted with more ordinary forms of free-associative thought. Much of what goes on under conventional short-term Ganzfeld conditions is almost certainly of the latter type, as indicated both by the character of typical mentation reports and by the EEG studies of Wackermann, Pütz, Büchi, Strauch, and Lehmann (2002). See also Alvarado (1998).
13. Our conception of this cosmogonic, world-generating, or virtual-reality system has much in common with that of Brann (1991), who provides a systematic exposition and defense of the thesis that an imaginative capacity of this sort constitutes a fundamental and neglected dispositional property of the mind.
14. See, in addition to those of Wallas (1926) and Hutchinson (1931, 1939), valuable works by Abrams (1958), Bowra (1949, 1955), Ghiselin (1952), Hadamard (1949), Harding (1948), Koestler (1964), Kubie (1958), and Prescott (1922).
15. It is of interest, and will be relevant to subsequent discussion, that Hadamard had read Myers and acknowledges (p. 22) that he himself was capable of automatic writing.
16. Myers, like Brann (1991), regards Blake as an example of strong imagination insufficiently controlled by supraliminal discipline: “Throughout all the work of William Blake (I should say) we see the subliminal self flashing for moments into unity, then smouldering again in a lurid and scattered glow” (HP, vol. 1, p. 73).
17. Similar group dynamics have sometimes been observed, incidentally, in the production of unusually strong psychokinetic effects (Batcheldor, 1984).
18. Two additional cases of this type were reported by Osty (1928a, 1928b), both involving pronounced aspects of automatism and trance.
19. We comment parenthetically that the conditions identified by Martindale are also associated with the hypnagogic state and its vivid, involuntary imagery, which commonly plays a role in creative activity as noted above. More generally, his neurobiological observations are also consistent with a Myers-like view in which global impairment of the normal supraliminal mode of mind-brain operation can lead to proportional disinhibition or release of normally inaccessible subliminal modes.
20. See especially Coleridge (1817/1967); good secondary sources here are Abrams (1958), Brann (1991), Hill (1978), and I. A. Richards (1960). It seems certain that Myers would have been intimately familiar with the original literature of this movement, and fully aware of its deep affinities with the theory of genius he articulates in HP. Most fundamentally, perhaps, for Coleridge as for Myers, the role of imagination in genius is secondary to its role as “the living Power and prime Agent of all human Perception” (Coleridge, 1817/1967, p. 167). For both, that is, perception becomes imagination constrained by sensory input. It is also somewhat ironic, even sad, that Coleridge is mainly cited by contemporary psychologists in connection with his much-maligned stories regarding the composition of Kubla Khan in an opium dream. He is of far greater importance to psychology, we think, as a theorist and exemplar of the imagination (see also Lowes, 1927).
21. Not all mainstream psychologists agree, however. R. J. Sternberg and Lubart (1999), for example, provide just a two-paragraph summary of psychoanalytic contributions, and then dismiss the entire subject from further consideration, saying only that “although the psycho-dynamic approach may have offered some insights into creativity, psychodynamic theory was not at the center of the emerging scientific psychology” (p. 6).
22. This same exercise, incidentally, was suggested by E. F. Kelly (1975, p. 72) as a way to strip the “semantic markers” of J. J. Katz and Fodor (1964) of their atmosphere of telegraphic speech.
23. On pp. 291–295 Hofstadter makes some especially interesting remarks concerning the position of Copycat’s computational architecture with respect to that of conventional symbolic and connectionist models. It is not a hybrid or mixture of these, but intermediate between the two. Most importantly, it embodies a fundamental operational distinction between types (the Platonic constituents of the slipnet) and tokens (their instances in the workspace), a distinction which Hofstadter believes is essential to all cognitive activity but cannot be captured by conventional connectionist models. His arguments here are in fact strikingly parallel to those of J. Fodor (2001), who has abandoned the CTM altogether.
24. It is a bit ironic, in this light, that despite his own earlier warnings about the Eliza effect in high-level analogy programs, Hofstadter himself begins talking ever more loosely about Copycat as “seeing,” “knowing,” “judging,” “understanding,” “focusing its attention,” “believing,” and so on. On the other hand, it is no surprise that he loathes the “biochauvin-ism” of philosophers such as John Searle (p. 290) and that he applauds the linguistic achievements of Terry Winograd’s SHRDLU without mentioning Winograd’s subsequent defection from AI (p. 311; see also our Chapter 1).
25. To describe this activity, of course, is not to explain it. Common to all recent attempts by cognitive psychologists to explain it, from Tversky (1977) to the real-world analogy theorists, is the idea that we can structure the representations of the relevant concepts and things in such a way that their “similarity” is intrinsic and can be directly computed from properties of those representations themselves. Formidable difficulties stand in the way of any such project, however, as shown in particular by philosophers such as McClendon (1955) and Goodman (1972). As Goodman summarizes his results, “similarity tends under analysis either to vanish entirely or to require for its explanation just what it purports to explain” (p. 446). For a notable recent attempt to find a way through these difficulties, see Goldstone (1994); but see also our Chapter 1 and below.
26. This is particularly true, we should point out, given the excessive latitude of Lombro-so’s criteria for “eccentricity.” These are sometimes bizarre by modern diagnostic standards, including things such as short stature, odd physiognomy or skull shape, left-handedness, leanness, rickets, and excessive yawning. His procedures also share with those of modern investigators such as Jamison (1993) a problem of circularity due to direct overlap between features of creativity itself and features supposedly diagnostic of the form of psychopathology with which it is being compared. For Lombroso, for example, “eccentricity” is also indicated by characteristics such as word-coining, originality, and flights of imagination.
27. Compare this statement by Charles Lamb from his remarkable three-page essay “Sanity of True Genius”: “The ground of the mistake is, that men, finding in the raptures of the higher poetry a condition of exaltation, to which they have no parallel in their own experience, besides the spurious resemblance of it in dreams and fevers, impute a state of dreaminess and fever to the poet. But the true poet dreams being awake. He is not possessed by his subject but has dominion over it” (E. V. Lucas, 1903, pp. 187–189).
28. Compare Barron (1958, p. 164). Much of the extensive research that Jamison cites in this connection derives from Frank Barron and the Institute for Personality Assessment and Research (IPAR). Interestingly enough, Barron (1969, preface) has revealed that his transition from philosophy to psychology was spurred by reading Myers’s Human Personality in combination with James’s Principles, and that this experience profoundly influenced all of his subsequent work.
29. Jamison (1993, pp. 105–113) identifies two principal characteristics which connect the thinking that occurs during this period with imagination as conceived by Myers, James, and Coleridge, among others: First, its extreme fluency, speed, and flexibility; second, its formal properties, including unusual forms of categorization and combination, merging of percepts, ideas, and images, highly original associations and analogies, and elevated mood. For her prime illustration of these characteristics, interestingly, she turns to Coleridge.
30. We do not mean to deny, however, that careless or superficial engagement with mediumship and associated practices such as the use of ouija boards and the like can sometimes result from or lead to psychological problems. We also wish to dissociate ourselves from most current commercial, stage, or TV “mediumship.”
31. Another aspect of this relationship, which we will not attempt to develop further here, concerns possible psi contributions to the creative process itself, and perhaps to some cases of “simultaneous discovery.” Mark Twain (1900), who was a firm believer in “mental telegraphy,” took special interest in this subject.
32. It is also of some importance here that Myers was socially connected to Wordsworth’s family, had met him as a child, and had previously written a monograph about him informed by intimate acquaintance with the poet’s family history and private correspondence (Myers, 1880/1929).
33. Treffert (1989) comes fairly close to Myers here, concluding that we must conceive the skills of prodigious savants as emerging out of some sort of “collective unconscious.” In such persons, Treffert says, “access to the rules of music or rules of mathematics, for example, is so extensive that some ancestral (inherited) memory must exist to account for that access” (p. 220). Unlike Myers, of course, Treffert hopes to locate this “ancestral memory” in the biology.
34. Myers held similar views in regard to music, which for him was a quintessential expression of subliminal faculty. “We know the difficulty of explaining its rise on any current theory of the evolution of human faculty. We know that it is like something discovered, not like something manufactured;—like wine found in a walled-up cellar, rather than like furniture made in the workshop above” (HP, vol. 1, p. 103). Mozart, for Myers as for Einstein, is the prime example. Although we will not pursue this subject further here, music remains even now resistant to conventional forms of explanation.
35. See also A. Baker (2005), for a philosophic defense based on the prime-numbered life-cycle lengths of cicadas. We specifically decline here to embrace any specific philosophical conception of the “ideal” realm, whether that of Plato himself or any of its historical variants. Like Myers and James, we tend to think that issues as to its ultimate character and composition will only be resolved, if ever, by expanded forms of empirical inquiry involving systematic exploration of altered states of consciousness. There is one point of Platonic doctrine, however, that is especially relevant to the subject of genius and deserves at least brief mention here. Myers was in general very ambivalent about the possibility of pre-existence, and he specifically rejected Plato’s conception of learning as “reminiscence”—the idea that Dase, for example, might have acquired his knowledge of the multiplication tables through “individual training” in some previous state of existence (HP, vol. 1, p. 119). We tend to agree with Myers in regard to Dase himself, but in light of the large body of evidence that has subsequently accumulated in support of reincarnation (see the Appendix), we would not rule out its possible contributions to some otherwise puzzling cases of extreme precocity. There are many such cases, possible examples including Mozart with his pronounced tendencies toward automatism, and Picasso, who drew skillfully before he could talk and whose first words were “piz, piz,” Spanish baby-talk for “pencil” (A. I. Miller, 2001, p. 10). Stevenson (2000, p. 654) has identified several instances of unusual precocity that occurred without familial counterparts, but we currently lack any such case that is accompanied by verifiable memories of an appropriate previous life.