The Sciences of Memory
I NOW WISH to advance four theses. They are difficult in themselves; their interconnections are yet more difficult. Here and in the next chapter I propose a way in which to understand the events I have been describing, both old and recent. Here are the theses, in capsule form.
1. The sciences of memory were new in the latter part of the nineteenth century, and with them came new kinds of truths-or-falsehoods, new kinds of facts, new objects of knowledge.
2. Memory, already regarded as a criterion of personal identity, became a scientific key to the soul, so that by investigating memory (to find out its facts) one would conquer the spiritual domain of the soul and replace it by a surrogate, knowledge about memory.
3. The facts that are discovered in this or that science of memory are a surface knowledge; beneath them is the depth knowledge, that there are facts about memory to be found out.
4. Subsequently, what would previously have been debates on the moral and spiritual plane took place at the level of factual knowledge. These political debates all presuppose and are made possible by this depth knowledge.
The idea of surface and depth knowledge is patterned after what Michel Foucault called connaissance and savoir. Foucault defined savoir as “a group of elements that would have to be formed by a discursive practice if a scientific discourse was to be constituted, specified not only by its form and rigor, but also by the objects with which it deals, the types of enunciation [roughly, statement] that it uses, the concepts that it manipulates, and the strategies that it employs.” As an example, he wrote that the savoir of psychiatry in the nineteenth century is not the sum of what was thought to be true, but “the whole set of practices, singularities, and deviations of which one could speak in psychiatric discourse.”1 Depth knowledge may not be known to anyone; it is more like a grammar, an underlying set of rules that determine, in this case, not what is grammatical, but what is up for grabs as true-or-false. Particular items counted as true, or as false, are connaissance, or what I call surface knowledge. My adjective “surface” is not intended to demean all our ordinary knowledge by implying that it is only on the surface, while there is something deeper that we ought to know. I pattern the terminology on Chomsky’s depth and surface grammar. Surface grammar is, for example, the grammar of English, which, you might say, is what matters. Some critics of Chomsky would say there is no such thing as depth grammar, and some critics of Foucault would say there is no such thing as his savoir. I use surface knowledge as an analytical idea, not to make a value judgment about kinds of knowledge.
This is not the place to substantiate my four theses for all the sciences of memory. There is a complex tale to tell about each one. Despite our deep programmatic commitment to the unity of science, there is not very much practical overlap between the sciences of memory. Think of (a) the neurological studies of the location of different types of memory; (b) experimental studies of recall; and (c) what might be called the psychodynamics of memory, which even Freud-haters can never entirely separate from Freud’s work. The word “dynamic” in psychology and psychiatry has had a checkered history.2 I mean the study of memory in terms of observed or conjectured psychological processes and forces.
All three of these sciences of memory are creatures of the nineteenth century. Only neurology has been deeply affected by high-technology advances in the twentieth century: we really can do things to brains of which nineteenth-century neurologists could only dream. To the three old sciences of memory we should add two twentieth-century branches of science. First, there is (d), work at the level of cell biology, transmission across potassium channels and the like. The ambition is certainly to join this together with (a), to provide an account, at the level of the cell and smaller, of the storage and transmission of information in different parts of the brain. Finally, we might add (e), computer modeling of memory in artificial intelligence, parallel distributive processing, and other branches of cognitive science.
These five kinds of science are connaissance, surface knowledge, that take for granted the objects they investigate. To call them surface is in no way to demean them. They matter in different ways. Some, whose present ratio of practical application to theoretical knowledge or speculation approaches one to infinity, may in the future change our daily lives. Funding agencies act on such hopes: there is nothing like a paragraph about Alzheimer’s disease to increase the probability of success for a grant application for work on ion currents across potassium channels. Nevertheless, the psychodynamics of memory is the only knowledge, of the three old sciences of memory that I have mentioned, that has profoundly influenced Western culture. Laboratory work on recall continues in a thousand departments of experimental psychology today. It has given us certain phrases of common speech—who does not know of short- and long-term memory? Yet its chief function, from a larger point of view, may be to shore up the depth knowledge, the conviction we do not state, that there is a body of facts about memory to be known.
I shall argue my four theses only in connection with (c), the psychodynamic approach to memory, which is, of course, a central aspect of therapy for multiple personality. But I would not want to fixate only on the ephemeral political battles of the moment, the brouhaha over false memory, for example. Memory has always had political or ideological overtones, but each epoch has found its own meaning in memory. Sometimes we can be quite nonplussed at what our predecessors have said. Let us take an example from my critical twelve years, 1874–1886. How could a lecture on memory perfectly enshrine the social pecking order of its day? On 12 July 1879 a talk to the Société de Biologie in Paris did exactly that.3 A Dr. Delannay told his audience that:
— People from the inferior races of modern times have better memories than those from the superior races. Blacks, Chinese, Italians, and Russians have a remarkable talent for learning languages (presumably, for learning French or English).
— The adult woman has a better memory than the man. Actresses learn their lines better and more quickly than actors. In undergraduate studies, female students do better than men.
— Adolescents have a better memory than adults. Memory is at its greatest powers at thirteen years of age and diminishes thereafter.
— The weak have a better memory than the strong. Memory is better among the less intelligent than the more intelligent. Children who get prizes for reciting from memory are less intelligent than others.
— The students at the Ecole Normale or at Val-de-Grâce—the school for military doctors—who have the best memories are not the most intelligent.
— Provincials have better memories than Parisians. Peasants have better memories than city-dwellers.
— Lawyers have better memories than doctors. Clerics have better memories than lay people.
— Musicians have better memories than other artists. One has better memory before eating than after. Education diminishes memory, in the sense that the illiterate have better memories than those who know how to write. One has a better memory in the morning than in the evening, in summer than in winter, in the south than in the north.
That pretty well covers the waterfront. Memory is an objective indicator of inferiority. An anticlerical physician has put priests and attorneys in their place, suitably ranked along with all the rest of humanity.
Delannay’s statement cheerfully combined the new sciences of memory and of anthropometry. Anthropometry—the name is due to Francis Galton (1822–1911)—was the measured and statistical part of anthropology. Anthropology was much occupied with comparisons between the different races of humankind, between subgroups within a region, and between the characteristics of the sexes. It generated measures of intelligence. Anthropology, sociology, and psychology were on the march, and part of the terrain they had to traverse was memory. This was the period when the sciences of memory came into being. The ideological bent of the nascent human sciences has been well chronicled, particularly in connection with racism and sexism. The political connotations of memory studies have not, however, been much noticed. But before we turn to these we should pause to confirm that the sciences of memory (a)–(c) were, in fact, new and not part of an old tradition.
One contrast between them and their predecessors is that between science and art, or between knowing that and knowing how. The new sciences of memory provided new knowledge that, as opposed to the art of memory, which taught us how to remember. No art was more carefully studied, or esteemed, from Plato until the Enlightenment, than the art of memory. Or perhaps we had better say the art of memorizing. This art was a collection of techniques or technologies of memory, variously called De arte memorativa, memoria technica, mnemonics.4 Plato and Aristotle refer often to one part of this art, particularly to a form of it that is translated as “placing.” A more helpful name is supplied by Mary Carruthers: architectural mnemonics.5 In the mind one forms the image of a three-dimensional space, a well-furnished house or even an entire city. Do you wish to remember that printing was invented in 1436? Then place a book in the thirty-sixth memory place in the fourth room of the first house in town. Cicero thought that such techniques, which survived long after the invention of printing, were of the highest importance, above all to the orator. Memory was also conceived of as essential to the formation of moral character; memory was highly ethical. The art of memory did languish until what are called the High Middle Ages. The greatest schoolmen, such as Thomas Aquinas, were marvels of memory. Carruthers argues for a complex relationship between books and memory; in many cases books were not the final, objective authorities that they later became, but mere adjuncts to the art of memory. The architectural mnemonic demanded rigorous discipline and regimens. One had to practice the building of houses and cities in the head, and learn how best to arrange things so that one could always be sure of where one had placed each object to be remembered. Texts were remembered in this way. Any competent scholar had an immense database stored in architectural mnemonics. Usually he could not go off to the library to check a citation or saying, but he had no need to do so. It was in his head.
Three things will be noticed. First, the art of memory had a central role in the ancient world, the High Middle Ages, and the Renaissance. Expertise in this art conferred great stature; it was a political asset. In the time of Cicero, it was an art for the orator, most esteemed of men. In the time of Aquinas, it was for the scholar. Carruthers makes a telling suggestion: “Memoria can be considered as one of the modalities of medieval culture (chivalry might be another).”6 It was, like chivalry, only for some, and its applications are limited to the highest pursuits. The ideological potential of “memory” was hardly an invention of 1879—only its content changed. Memory was for the elite, and yet, like chivalry, it permeated the world. Memoria, Carruthers continues, “is also a value in itself, identified with the virtue of prudence. As modalities, values enable certain behavior, and also give greater privilege to some behavior over others.”
Second, the art of memory was truly a techne, a knowing how, and not a knowing that. It was not a science that delivered knowledge about some object of study, “the memory.” Third, the art of memory is outer-directed. It is at most incidentally concerned with remembering one’s own experiences. The whole point is to provide instant recall of any body of desired facts, things, or texts. One arranges external material in a vivid picture in one’s mind, to which one has direct access. Perhaps what we call computer memory, with its numerous technologies, is the lineal descendant of the art of memory. There is something linguistically adventitious about this. Every language carves the memory ideas into different groupings of words. In German neither Erinerrung nor Gedächtnis would serve at all for the memory of a computer, so the word is simply Speicher, storehouse. Medievals commonly used the metaphor of the storehouse for memory.
The art of memory waned during the Enlightenment, but it was not replaced by another art or science. Mnemonics were still taught yet were not invested with any moral authority or stature. Of course people did not lose interest in memory. One of the most moving statements about memory and its recovery—flashbacks, even—was penned by a most unlikely author, John Locke:
The Mind very often sets it self on work in search of some hidden Idea, and turns, as it were, the Eye of the Soul upon it; though sometimes too they start up in our Minds of their own accord, and offer themselves to the Understanding; and very often are rouzed and tumbled out of their dark Cells into open Day-light, by some turbulent and tempestuous Passion; our affections bringing Ideas to our Memory, which had otherwise lain quiet and unregarded.7
In Locke’s day there was no systematic attempt to uncover facts about memory. That began only late in the nineteenth century. Of course for every predecessor there is a predecessor. The localization project of neurology derives in part from phrenology, which located mental faculties and abilities by means of bumps on the skull. But only in 1861 did an anatomist open up a brain and identify a lesion with the loss of a mental faculty. That was Paul Broca (1824–1880). “We have every reason to believe that, in this case, the lesion of the frontal lobe was the cause of the loss of language.”8 (It will be recalled from chapter 11 that three years earlier Broca had enthusiastically tried out Azam’s hypnotism in an actual surgical operation on an abscess.) Broca continued his work on localization until his death, but he was also enormously active in French anthropology, which was, in the first instance, very much a study of race and races. We remember him for Broca’s region, the motor speech center of the brain. Broca successfully began the great neurological program, still with us, of locating different faculties in different parts of the brain. Broca’s discovery generated enthusiastic research. Historians find the next landmark in Carl Wernicke’s identification of another region in which words (or word images) are stored. This could be regarded as the first delineation of a part of the brain that serves as a specific type of memory bank. If a single essay pulled all this together, it was Ludwig Lichtheim’s 1885 study of aphasia.9 I should emphasize that this is an anatomical, physiological program, which we call neurological because the part of the body that is examined is the brain.
Now let us turn to the second science of memory, namely, recall. In 1879 Hermann Ebbinghaus (1850–1909) established a new paradigm for psychological research. It was far from the first experimental psychology. For example, Gustav Fechner’s psychophysics transformed the experimental investigation of the relationship between body and mind. Fechner (1801–1887) discovered empirical laws about the least differences in pairs of weights that could be discerned by an experimental subject. There was experimentation in Germany before Fechner, and a great deal after. Nevertheless, Kurt Danziger plausibly takes Ebbinghaus to have inaugurated psychology as a laboratory science of measurement. “All the fundamental features of the measurement of psychological capacity were first manifested in Hermann Ebbinghaus’s classical work on memory.”10 Ebbinghaus’s research became public in 1885 with a major book, On Memory.11
Ebbinghaus wanted to study memory in its pure form, uncontaminated by other kinds of knowledge. So he experimented on the recall of nonsense syllables. Why is this so important? David Murray asserts that G. E. Müller (1850–1934) was far more influential, because he pioneered the interference theory of forgetting, and because Ebbinghaus himself was so entirely empirical, not speculating on the mechanisms of memory.12 Why then has Danziger singled out Ebbinghaus as a “first,” comparable to Broca? Aside from the great revolutionaries in the sciences, “firsts” are picked not so much for the importance of their contribution as for the way that they conveniently mark, for us, a new departure. The critical feature of Ebbinghaus’s work was that he instituted statistical treatment of data. Memory was to be investigated in the context of the subject’s ability to recall a series of nonsense syllables. Then one was to construct a statistical analysis of the ability to recall. Ebbinghaus began work on himself, a typical human being, but his behavior was to be understood only through statistical scrutiny. His approach became standard, integrated with learning theory. Whole cohorts of research psychologists have devoted their entire careers to continuing in the footsteps of Ebbinghaus. Few journals of experimental psychology will even consider refereeing a research paper that does not include a battery of statistical tests. Here, then, we have a remarkable conjunction: the first sustained study of recall and the first sustained use of statistical analysis in psychology.13 If Broca conveniently marks the start of the anatomical science of memory, Ebbinghaus conveniently marks the start of the statistical science of memory.
For my purposes, the anatomical and statistical studies of memory are only asides, which is why I hang them on standard historical pegs like Broca and Ebbinghaus. In contrast, we have been immersed in psychodynamics from the start, and as soon as you go into detail there are no firsts. Instead I shall choose a figure who serves as an ideal type—one of many who well displays the third new science of memory. In 1879 Théodule Ribot, in Paris, gave a set of lectures on the diseases of memory. They were published in 1881, the first of a trio of books on diseases—of memory, of the will (1883), and of the personality (1885).14 Coincidences abound: Ribot began lecturing on the topics that formed this sequence in Paris, in 1879, the year that Ebbinghaus, in Leipzig, began his memory experiments. He completed the trio in 1885, the year that Ebbinghaus published his results, and the year that Lichtheim brought together the state of the art on localization of brain function, including memory of words. Annual coincidences mean nothing in themselves, but we may begin to get a picture of three relatively unconnected sciences of memory driving on at about the same time, and at about the same pace.
The sciences take different courses in different institutional settings, and in different cultural or national environments. The development of psychology in France was very different from that in Germany or America. The French route was medical and pathological.15 As a result the Parisian study of memory was the study of forgetting. Michael Roth has written elegantly about deeper French cultural meanings in the medical fascination with forgetting and nostalgia.16 He notes that although most of Ribot’s book is about forgetting, it concludes with a curious chapter on hypermnesia, too much memory, which was thought to be pathological. So he sees Ribot’s text as almost a moral tract, intent to define the amount of memory that is just right.
Roth’s analysis of the book’s subtext is insightful, but more mundane facts of institutional history should also be taken into account. Danziger opens his book with a striking insight: In Germany and America experimental psychology patterned itself on experimental physiology; it was even called “physiological psychology.”17 The situation was entirely different in France. Psychiatry had always been a major feature of French medicine, ever since Pinel “liberated the asylums” at the end of the eighteenth century. The charismatic influence of the neurologist Charcot, from early in the 1870s until his death in 1893, made the connections between mind, brain, and mental illness central to scientific study. Hence when one took up the psychological study of memory in Paris, in the 1870s, one could hardly fail to start with its pathologies, with forgetting and amnesia.
The effect was not entirely confined to France. The United States was in those days eclectically open to all new scientific ideas from anywhere in Europe. The article “Memory” in Baldwin’s classic Dictionary of Philosophy and Psychology (1901) is only half as long as the article “Memory, defects of.” The latter chiefly focuses on amnesia. The word “amnesia,” or rather amnésie, was used in French in 1771, in the translation from the Latin of Sauvages’s immensely influential nosology, or classification, of all diseases.18 From the beginning the word named a medical disorder, a potential object of knowledge. But it was not an important research field until the 1870s. Then it became central to the new French science of memory.
As my “ideal type” for this new science I wanted to choose a figure who was not a pathologist or neurologist, and I wanted someone who was prepared not only to state facts but also to discuss method. That is why I chose Ribot, by training a philosopher. Since I single him out, I must make plain that his positive views about memory (as opposed to forgetting) were trite. He was a loyal disciple of British associationist psychology, acknowledging his debt to Scottish authors on the first page of his book.19 He usefully insisted that we should not be talking about memory, as if there were just one faculty, but about “memories” (mémoires). But this is only a deduction from the claims that different types of acquired abilities, skills, and knowledge are stored in different parts of the brain. On the relations between mind and brain, Ribot was no more, and no less, programmatic than most other positivist or scientistic writers of the day. “Memory,” he wrote, “is essentially a biological fact, accidentally a psychological fact.”20 He took the unconscious (l’inconscient) very seriously, and in ways very different from Eduard von Hartmann’s massive and massively romantic Philosophy of the Unconscious of 1869.21 But he did so only as part of his purely speculative neurophysiology. Consciousness comprised certain events in the nervous system (especially “discharges,” in the parlance of the day) that endured more than a certain finite time. Events of the same type, but briefer, were unconscious. “The brain is like a laboratory full of movement, where thousands of tasks are performed at once. Unconscious cerebration, not being subject to the condition of time, takes place, so to speak, only in space, and may act in several places at once. Consciousness is the narrow gate through which a very small part of this work appears to us.”22 Such talk of the unconscious was so common in his day that it would be foolish to see Ribot as anticipating Janet’s idea of the subconscious, or Janet’s word sous-conscience. Janet himself used the word inconscient in the psychological essays preceding his Psychological Automatism of 1889. At that point he coined the name “subconscious” to separate himself from the tradition of Hartmann that still persisted in Germany.
Ribot held the chair of experimental and comparative psychology at the Collège de France. Recall how Pierre Janet, Ribot’s successor in this chair, said (with some exaggeration), “But for Félida, it is not certain that there would be a professorship at the Collège de France and that I should be here speaking to you of the mental state of hystericals.” In chapter 11 I said a little about French positivism, as argued by powerful cultural leaders such as Hippolyte Taine and Emile Littré. Theirs was the 1870s model, popular as a response to inglorious defeat by Prussia; it was republican and secular. Ribot was forthright in subscribing to their school of thought. He subtitled his 1881 book on memory An Essay in the Positive Psychology. In that book he discussed “the detailed and instructive observations of Dr. Azam.” After describing work on dédoublement, he wrote:
Let us first reject the idea of a moi conceived as an entity distinct from states of consciousness. That is a useless and contradictory hypothesis; it is an explanation worthy of a psychology in its infancy, which takes as simple that which appears simple, and which postulates instead of explaining. I join in the opinion of contemporaries who see the conscious person as a compound, a resultant of very complex states.23
Ribot went on to explain that there are two ways to consider the moi. As the moi appears to itself, it is a collection of present states of consciousness and may be compared to a present visual field. But “this moi of each moment, this present perpetually renewed is in large part fed by memory…. In a word, the moi can be considered in two ways: either in its actual form, and then it is the sum of actual states of consciousness; or in its continuity with its past, and then it is formed by memory.”24
Ribot began his next book but one, on diseases of the personality, by saying, “It is but natural that the representatives of the old school, slightly bewildered at the situation [in psychology], should accuse the adherents of the new school of ‘stealing their moi.’”25 The “old school” was, as explained in chapter 11, the so-called eclectic spiritualism of Victor Cousin. The strategy of Ribot and his positivist colleagues was not to attack religious or philosophical ideas of the soul, but to provide a surrogate for the one aspect of a human being that seemed resistant to science. Instead of studying a unitary moi, we should study memory. But how do we know that there is no unitary self? The cases of dédoublement, Félida and her successors, seemed splendid for showing that a person was not constituted by a single transcendental, metaphysical or spiritual self or ego. For in those individuals, there was not one single self. Those individuals had two personalities, each connected by a continuous or normal chain of memories, aside from amnesic gaps. At least one personality was ignorant of the other. Hence (it seemed) there were two persons, two souls in one body.
The use of doubles to refute the idea of a transcendental ego was more rhetorical than logical. The rhetoric succeeded by changing the ground on which to think about the soul. The soul was the last bastion of thought free of scientific scrutiny. To be sure, there had long been mechanical models of the human being, including such as that presented in La Mettrie’s scandalous book published in Holland in 1747, Man a Machine. The French positivists undoubtedly believed that all psychology would in the end have a neurological foundation. That was a commonplace, shared, for example, by Freud and by many of his German-speaking predecessors. The importance of Ribot and his peers was not that they had a program but that they offered knowledge. It was new knowledge, scientific knowledge about memory. Real knowledge, scientific laws about memory, even what is still called “Ribot’s law.” This law is a perfect example of surface knowledge, a statement about how the memory faculties decay, presupposing that those faculties are objects of a certain sort. His own name for the law was the law of regression or reversion. “The progressive destruction of the memory,” owing to whatever pathology, “follows a logical order, a law. It advances progressively from the unstable to the stable.” Memories and skills acquired early are the stable ones, while the more recent are more unstable. His evidence was taken from various types of amnesia, including traumatic amnesia and dementia in senility. He held his law to be universal, to be applicable to any type of memory loss. His law seemed to him “to follow from facts, and to demand recognition as an objective truth.”26 I consign his own statement of the law to a note.27 Our concern is the kind of law that it purports to be. It is an objective truth. It follows from facts. The facts in question are from pathological psychiatry. It is a law about loss of memory, about forgetting. Finally, the law covers, in a uniform way, both forgetting caused by physical lesions, and forgetting caused by psychic shock. Thus it covers trauma in the old sense of the word, and trauma in the about-to-be (this is 1881) sense of the word. When you stand back, ignoring the content of Ribot’s law and looking only at its form, you see that it foreshadows the form of almost all subsequent dynamic psychiatry.
I am not saying that Ribot is a precursor of Freud, the modern multiple movement, or whatever. I am saying that he is an early instance of a man whose surface knowledge is worked out within the rules of that underlying depth knowledge which remains the depth knowledge to this very day. One feature of the modern sensibility is dazzling in its implausibility: the idea that what has been forgotten is what forms our character, our personality, our soul. Where did we get that idea? To grasp this we need to reflect on how knowledge about memory became possible late in the nineteenth century. What were the new sciences of memory trying to do? Find out, of course, and more power to them. But although I have argued the case for only one of the new sciences, I suggest that they all emerged as surrogate sciences of the soul, empirical sciences, positive sciences that would provide new kinds of knowledge in terms of which to cure, help, and control the one aspect of human beings that had hitherto been outside science. If we address only the surface facts about memory, the politicization of memory will seem only a curious accident. But if we think of how the very idea of such facts came into being, the battles may seem almost inevitable.