Memory and Imagination
By submitting to the primacy of the question “What?” the phenomenology of memory finds itself at the outset confronting a formidable aporia present in ordinary language: the presence in which the representation of the past seems to consist does indeed appear to be that of an image. We say interchangeably that we represent a past event to ourselves or that we have an image of it, an image that can be either quasi visual or auditory. In addition to ordinary language, a long philosophical tradition, which surprisingly combines the influence of English-language empiricism with the rationalism of a Cartesian stamp, considers memory the province of the imagination, the latter having long been treated with suspicion, as we see in Montaigne and Pascal. This continues to be the case, most significantly, in Spinoza. We read in proposition 18 of the second part of the Ethics: On the Nature and the Origin of the Soul: “If the human Body has once been affected by two or more bodies at the same time, then when the Mind subsequently imagines one of them, it will immediately recollect the others also.”1 This sort of short-circuit between memory and imagination is placed under the sign of the association of ideas: if these two affections are tied by contiguity, to evoke one—to imagine it—is to evoke the other—to remember it. Memory, reduced to recall, thus operates in the wake of the imagination. Imagination, considered in itself, is located at the lowest rung of the ladder of modes of knowledge, belonging to the affections that are subject to the connection governing things external to the human body, as underscored by the scholia that follows: “this connection happens according to the order and connection of the affections of the human Body in order to distinguish it from the connection of ideas which happens according to the order of the intellect” (466). This declaration is all the more remarkable in that we read in Spinoza a magnificent definition of time, or rather of duration, as “continuation of existence.” What is surprising is that memory is not related to this apprehension of time. And as memory, considered, moreover, as a mode of learning, in terms of the memorization of traditional texts, has a bad reputation—see Descartes’s Discourse on Method—nothing comes to the aid of memory as the specific function of accessing the past.
As a countercurrent to this tradition of devaluing memory, in the margins of a critique of imagination, there has to be an uncoupling of imagination from memory, as far as this operation can be extended. The guiding idea in this regard is the eidetic difference, so to speak, between two aims, two intentionalities: the first, that of imagination, directed toward the fantastic, the fictional, the unreal, the possible, the utopian, and the other, that of memory, directed toward prior reality, priority constituting the temporal mark par excellence of the “thing remembered,” of the “remembered” as such.
The difficulties inherent in this operation of uncoupling hearken back to the Greek origin of the problematic (section 1: “The Greek Heritage”). On the one hand, the Platonic theory of the eikōn places the main emphasis on the phenomenon of the presence of an absent thing, the reference to past time remaining implicit. This problematic of the eikōn has its own relevance and its own proper instance, as our subsequent investigations will confirm. Nevertheless, it has become an obstacle to recognizing the specificity of the properly temporalizing function of memory. We must turn to Aristotle to find an acknowledgment of this specificity. The proud declaration that we read in the magnificent little text of the Parva naturalia: On Memory and Recollection—“All memory is of the past”—will become our lodestar for the rest of our exploration.
The central part of this study, “A Phenomenological Sketch of Memory,” will be devoted to an effort to form a typology of mnemonic phenomena. Despite its apparent dispersion, this study aims at determining the original experience of temporal distance, of the depth of time past, through a series of approximations. I will not conceal the fact that this plea on behalf of memory’s mark of distinction has to be paired with a parallel revision of the thematic of the imaginary, similar to what Sartre undertook to do in his two books, L’Imagination and L’Imaginaire, a revision that would tend to dislodge the image from its alleged place “in” consciousness. The critique of the picture-image would then become one document in the file common to imagination and to memory, a file that begins with the Platonic theme of the presence of the absent.
However, I do not think that one can be content with this twofold operation of specifying the imaginary, on one hand, and memories, on the other. There must be an irreducible feature in the living experience of memory that explains the persistence of the confusion conveyed by the expression “memory-image.” It does appear that the return of a memory can only take place in the mode of becoming-an-image. The parallel revision of the phenomenology of memories and the phenomenology of images will encounter its limit in this image-making process of memories (section 3: “Memories and Images”).
The constant danger of confusing remembering and imagining, resulting from memories becoming images in this way, affects the goal of faithfulness corresponding to the truth claim of memory. And yet . . .
And yet, we have nothing better than memory to guarantee that something has taken place before we call to mind a memory of it. Historiography itself, let us already say, will not succeed in setting aside the continually derided and continually reasserted conviction that the final referent of memory remains the past, whatever the pastness of the past may signify.
§
The problem posed by the entanglement of memory and imagination is as old as Western philosophy. Socratic philosophy bequeathed to us two rival and complementary topoi on this subject, one Platonic, the other Aristotelian. The first, centered on the theme of the eikōn, speaks of the present representation of an absent thing; it argues implicitly for enclosing the problematic of memory within that of imagination. The second, centered on the theme of the representation of a thing formerly perceived, acquired, or learned, argues for including the problematic of the image within that of remembering. These are the two versions of the aporia of imagination and memory from which we can never completely extricate ourselves.
Plato: The Present Representation of an Absent Thing
It is important to note from the start that it is within the framework of the dialogues on the sophist and, through this person, on sophistry itself and the properly ontological possibility of error, that the notion of the eikōn is encountered, either alone or paired with that of the phantasma. In this way, from the very outset, the image but also by implication memory are cast under a cloud of suspicion due to the philosophical environment in which they are examined. How, asks Socrates, is the sophist possible and, with him, the false-speaking and, finally, the non-being implied by the non-true? It is within this framework that the two dialogues bearing the titles Theaetetus and Sophist pose the problem. To complicate matters further, the problematic of the eikōn is, in addition, from the outset associated with the imprint, the tupos, through the metaphor of the slab of wax, error being assimilated either to an erasing of marks, semeia, or to a mistake akin to that of someone placing his feet in the wrong footprints. We see by this how from the beginning the problem of forgetting is posed, and even twice posed, as the effacement of traces and as a defect in the adjustment of the present image to the imprint left as if by a seal in wax. It is noteworthy that memory and imagination already share the same fate in these founding texts. This initial formulation of the problem makes all the more remarkable Aristotle’s statement that “all memory is of the past.”
Let us reread the Theaetetus, beginning at 163d.2 We are at the heart of a discussion centered around the possibility of false judgment, which concludes with a reference to the thesis that “knowledge is simply perception” (151e–187b).3 Socrates proposes the following “attack”: “Supposing you were asked, ‘If a man has once come to know a certain thing, and continues to preserve the memory of it, is it possible that, at the moment when he remembers it, he doesn’t know this thing that he is remembering?’ But I am being long-winded, I’m afraid. What I am trying to ask is, ‘Can a man who has learned something not know it when he is remembering it?’” (163d). The strong tie of the entire problematic to eristic is immediately obvious. Indeed, it is only after having crossed through the lengthy apology of Protagoras, and his open pleading in favor of the measure of man, that a solution begins to dawn, but, before that, an even more pointed question is raised: “Now, to begin, do you expect someone to grant you that a man’s present memory of something which he has experienced in the past but is no longer experiencing is the same sort of experience as he then had? This is very far from being true” (166b). An insidious question, which leads the entire problematic into what will appear to us to be a trap, namely, resorting to the category of similarity to resolve the enigma of the presence of the absent, an enigma common to imagination and memory. Protagoras tried to enclose the authentic aporia of memories, namely, the presence of the absent, in the eristic of the (present) non-knowledge of (past) knowledge. Armed with a new confidence in thinking, likened to a dialogue of the soul with itself, Socrates develops a sort of phenomenology of mistakes, where one thing is taken for another. To resolve this paradox he proposes the metaphor of the block of wax: “Now I want you to suppose, for the sake of the argument, that we have in our souls a block of wax, larger in one person, smaller in another, and of pure wax in one case, dirtier in another; in some men rather hard, in others rather soft, while in some it is of just the proper consistency.” Theaetetus: “All right, I’m supposing that.” Socrates: “We may look upon it, then, as a gift of Memory [Mnemosyne], the mother of the Muses. We make impressions upon this of everything we wish to remember [mnēmoneusai] among the things we have seen or heard or thought of ourselves; we hold the wax under our perceptions and thoughts and take a stamp from them, in the way in which we take the imprints [marks, sēmeia] of signet rings. Whatever is impressed upon the wax we remember and know so long as the image [eidōlon] remains in the wax; whatever is obliterated or cannot be impressed, we forget [epilelēsthai] and do not know” (191d). Let us note that the metaphor of the wax conjoins the problematics of memory and forgetting. There follows a subtle typology of all the possible combinations between the moment of knowledge and the moment of the acquisition of the imprint. Among these, let us note the following pairs: “that a thing which you both know and are perceiving, and the record of which you are keeping in its true line [ekh ōn to mnēmeion orthōs] is another thing which you know . . . that a thing you both know and are perceiving and of which you have the record correctly in line as before, is another thing you are perceiving” (192b–c). It is in an effort to identify this veridical characteristic of faithfulness that we will later reorient the entire discussion. Pursuing the analogy of the imprint, Socrates assimilates true opinion to an exact fit and false opinion to a bad match: “Now, when perception is present to one of the imprints but not to the other; when (in other words) the mind applies the imprint of the absent perception to the perception that is present; the mind is deceived in every such instance” (194a).4 We need not linger over the enumeration of the different kinds of wax, intended as a guide to the typology of good or bad memories. But let me not fail to mention, however, for our reading pleasure, the ironic reference (194e–195a) to “those whose wax is shaggy” (Iliad II!) and “soft.” Let us retain the more substantive idea that false opinion resides “not in the relations of perceptions to one another, or of thoughts to one another, but in the connecting [sunapsis] of perception with thought” (195c–d). The reference to time we might expect from the use of the verb “to preserve in memory” is not relevant in the framework of an epistemic theory that is concerned with the status of false opinion, hence with judgment and not with memory as such. Its strength is to embrace in full, from the perspective of a phenomenology of mistakes, the aporia of the presence of absence.5
With regard to its impact on the theory of imagination and of memory, it is the same overarching problematic that is responsible for the shift in metaphor with the allegory of the dovecote.6 Following this new model (“the model of the aviary” in the words of Burnyeat), we are asked to accept the identification between possessing knowledge and actively using it, in the manner in which holding a bird in the hand differs from keeping it in a cage. In this way, we have moved from the apparently passive metaphor of the imprint left by a seal to a metaphor that stresses power or capacity in the definition of knowledge. The epistemic question is this: does the distinction between a capacity and its exercise make it conceivable that one can judge that something one has learned and whose knowledge one possesses (the birds that someone keeps) is something that one knows (the bird one grabs in the cage) (197b–c)? The question touches our discussion inasmuch as a faulty memorization of the rules leads to an error in counting. At first glance, we are far from the instances of errors of fit corresponding to the model of the block of wax. Were these not, nevertheless, comparable to the erroneous use of a capacity and, by this, to a mistake? Had not the imprints to be memorized in order to enter into use in the case of acquired knowledge? In this way the problem of memory is indirectly concerned by what could be considered a phenomenology of mistakes. The failed fit and the faulty grasp are two figures of mistakes. The “model of the aviary” is especially well-suited to our investigation inasmuch as grasping is in every case comparable to a possession (hexis or ktēsis), and above all to hunting, and in which every memory search is also a hunt. Let us again follow Socrates, when, as a true sophist, he surpasses himself in subtleties, mixing ring doves with doves but also non-doves with real doves. Confusion is rampant not only at the moment of capture but also with respect to the state of possession.7
By these unexpected divisions and duplications, the analogy of the dovecote (or the model of the aviary) reveals a richness comparable to that of the foot mistakenly placed in the wrong print. To the mis-fit is added the erroneous grasp, the mis-take. However, the fate of the eikōn has been lost from sight. The Sophist will lead us back to it.
The problematic of the eikōn developed in the Sophist comes directly to the aid of the enigma of the presence of absence concentrated in the passage in Theaetetus 194 related above.8 What is at stake is the status of the moment of recollection, treated as the recognition of an imprint. The possibility of falsehood is inscribed in this paradox.9
Let us focus on the key passage in the Sophist in which Plato distinguishes veracity from trickery in the order of imitation (234ff.).10 The framework of the discussion resembles that of the Theaetatus: how are sophistry and its art of illusion possible? The Stranger and Theaetatus are in agreement in saying that the sophist—him again!—is principally an imitator of being and of truth, someone who manufactures “imitations” (mimēmata) and “homonyms” (homōnuma) of beings (234b). Here we change metaphors. We pass from the imprint in wax to the portrait, the metaphor extending in its turn from graphic arts to language arts (eidōla legomena, “spoken copies of everything” capable of making us believe “the words are true” [234c]). We are thus in the midst of technique, of mimetic technique, in which imitation and magic (“a kind of magician” [235b5]) are indistinguishable. Within this assigned framework Plato practices his favored method of division: “We’ll divide the craft of copymaking [eidōlopoiikēn tekhnēn] as quickly as we can” (235b). On one side we have tekhnē eikastikē (“the art of likeness-making. That’s the one we have whenever someone produces an imitation by keeping to the proportions of length, breadth, and depth of his model, and also by keeping to the appropriate colors of its parts” [235d–e]). On the other side we have the simulacrum or appearance, for which Plato reserves the term phantasma. So here we have eikōn opposed to phantasma, “eikastic” art to “fantastic” art, the making of likenesses to the making of appearances (236c). With regard to its specific character, the problem of memory has disappeared, overwhelmed by the dominant problematic, namely, the question of knowing in what compartment the Sophist can be placed. The Stranger confesses his bafflement. The entire problem of mimetics is, by the same stroke, dragged into the aporia. To get out of it, it will be necessary to move higher in the hierarchy of concepts and assume nonbeing.
The idea of “faithful resemblance” belonging to the eikastic art will at least have served as a relay. Plato seems to have noted the threshold of the impasse, when he asks himself: “what in the world do we mean by a ‘copy’ [eidōlon]?” (239d). We lose our way in the enumeration of examples that seem to escape the art of orderly division and, first of all, that of generic definition: “What in the world would we say a copy is, sir, except something that’s made similar (heteron) to a true thing and is another thing that’s like it?” (240a). But what is the meaning of “a true thing”? And “another thing”? And “like it”? Now we are at sea: “So you’re saying that which is like [eikōna] is not really that which is, if you speak of it as not real existence, if you are going to call it not true?” (240b). To say this, we recognize that we have been forced “to agree unwillingly that that which is not in a way is” (240c). The phenomenological difference, as it were, between eikastic and fantastic is caught up in the whirlwind in which eristic and dialectic are scarcely distinguishable. All of this is perhaps due to the fact that the sophist’s question of being has swamped the discussion, and the battle against Parmenides—against “what our Father says” (241d)—has expended all the intellectual energy. We even see the three terms, eidōlon, eikōn, and phantasia, reunited under the ignominious charge of deception (apatō, 260c), and a bit later: “copy-making and appearance-making [eidōlopoiikēn kai phantastikēn]” (260d). It is simply recommended that “we have to search around for the nature of speech [logos], belief [doxa], and appearance [phantasia]” (260e) from the viewpoint of their “association with that which is not” (ibid.).
Let us take stock of the aporetic results of our passage through the Platonic texts on memory. We can lay out the difficulties in the following order. The first has to do with the absence (noted in passing) of explicit reference to the distinctive feature of memory, namely, the anteriority of “marks,” sēmeia, in which the affections of the body and the soul to which memory is attached are signified. It is true that on many occasions past verb tenses are explicitly employed, but there is no separate reflection devoted to these indisputable deictic forms. It is on this point that Aristotle’s analyses will mark a clear break.
The second difficulty concerns the sort of relation that exists between the eikōn and the first mark, as this is sketched out within the framework of the imitative arts. To be sure, the distinction made in the Sophist between eikastic art and fantastic art is vigorously affirmed. And we can consider this distinction to be the starting point for a full recognition of the problematic at the center of this study, namely, the truthful dimension of memory and, let us add in anticipation, of history. Moreover, throughout the debate over sophistry, the epistemological and ontological status given to falsity presupposes the possibility of wresting true discourse away from the vertigo of falsity and its real nonbeing. The chances of a true icon are, therefore, preserved. But if the problem is recognized in its specificity, the question arises whether the requirement of faithfulness, of veracity, contained in the notion of an eikastic art finds an appropriate framework within the notion of a mimetic art. The result of this classification is that the relation to the signifying marks can only be a relation of similarity. In Time and Narrative, vol. 1, I explored the resources of the concept of mimēsis and attempted to give it a wider scope at the cost of deepening the split between mimēsis and imitation as copy. The question nevertheless remains whether the problematic of similarity does not constitute a diriment impediment to recognizing the specific features that distinguish memory from imagination. Can the relation to the past be only a variety of mimēsis? This difficulty will continue to hound us. If our doubt is well founded, the idea of “faithful resemblance,” proper to the eikastic art, will likely turn out to be a mask rather than a way station in the exploration of the truthful dimension of memory.
We have not yet reached the end of the impasse. We saw the Theaetetus link the study of the eikōn closely to the assumption of a mark comparable to the imprint of a seal upon a block of wax. We recall the terms in which the Theaetetus made the connection between eikōn and tupos: “Suppose, for the sake of argument, that we have in our souls a block of wax . . .” (191c). This assumption is supposed to allow us to solve the puzzle of confusing or mistaking, not to mention the puzzle of the persistence of the marks or, again, that of their effacement in the case of forgetting. This gives an indication of the burden the hypothesis bears. In this regard, Plato does not hesitate to place the hypothesis under the sign of Mnemosyne, the mother of all the Muses, thereby lending a tone of solemnity to the hypothesis. The alleged conjunction of eikōn and imprint is thus held to be more primitive than the relation of resemblance that sets the mimetic art into play. Or, to say this in other words, there can be a truthful or deceitful mimetic because there is between the eikōn and the imprint a dialectic of accommodation, harmonization, or adjustment that can succeed or fail. With the problematic of the imprint and that of the relation between eikōn and imprint, we have reached the end-point of the entire regressive analysis. This hypothesis—or better, admission—of the imprint has, over the course of the history of ideas, produced a procession of difficulties that have continued to overwhelm not only the theory of memory, but also the theory of history, under another name—the “trace.” History, according to Marc Bloch, aspires to be a science of traces. It is now possible to lift some of the confusion relative to the use of the word “trace” in the wake of the term “imprint.” Applying the Platonic method of division recommended—and practiced—by Plato in the Sophist, I distinguish three major uses of the word “trace.”
I provisionally set aside the traces on which historians work: these are traces that are written and eventually archived. These are the ones Plato has in mind in the myth of the Phaedrus that recounts the invention of writing. We shall return to them in the prelude to part 2. A dividing line will thus be drawn between the “external” marks of writing properly speaking, of written discourse, and the graphic component inseparable from the eikastic component of the image by virtue of the metaphor of the wax impression. The myth of the Phaedrus will tip the typographical model, upon which David Farrell Krell constructs his interpretation of the Theaetetus, from the intimacy of the soul to the exteriority of the public writing of discourse. The origin of written traces will become for all that more mysterious.
Quite another matter is the impression as an affection resulting from the shock of an event that can be said to be striking, marking. This impression is essentially undergone, experienced. It is tacitly presupposed in the very metaphor of the tupos at the moment the seal is pressed into the wax, inasmuch as it is the soul that receives the imprint (Theaetetus 194c). It is explicitly invoked in the third Platonic text which we will now consider. This text is found in the Philebus 38a–39c.11 Once again, it is opinion, sometimes true, sometimes false, that is at issue here, in this case in its relation to pleasure and pain, initial candidates in the competition among rival goods presented at the beginning of the dialogue. Socrates proposes: “And is it not memory and perception that lead to judgment or the attempt to come to a definite judgment as the case may be?” (38b). Protarchus acquiesces. Then comes the example of someone who wants to “distinguish” (krinein) what appears to him from afar to be a man. What happens when it is to himself that he addresses his questions? Socrates proposes: “That our soul in such a situation is comparable to a book” (38e). “How so?” asks Protarchus. The explanation follows: “If memory and perceptions concur with other impressions [pathēmata] at a particular occasion, then they seem to inscribe [graphein] words in our soul, as it were. And if what [the experience, pathēma] is written is true, then we form a true judgment and a true account of the matter. But if what our scribe [grammateus] writes is false, then the result will be the opposite of the truth” (39a).12 And Socrates then proposes another comparison, this time with painting, a variant of graphism: “Do you also accept that there is another craftsman [dēmiourgos] at work in our soul at the same time?” Who? “A painter [zōgraphos], who follows the scribe and provides illustrations [graphei] to his words” (39b). This takes place as a result of a separation between, on the one hand, opinions and discourses accompanied by sensation and, on the other hand, “the images he has formed inside himself” (39b). Such is the inscription in the soul to which the Phaedrus will oppose the external marks with which written discourse is constructed. The question posed by this affection-impression is therefore twofold: on the one hand, how is it preserved, how does it persist, whether or not it is recalled? On the other, what meaningful relation does it maintain in relation to the marking event (what Plato calls eidōlon to avoid confusion with the present eikōn of the absent mark that poses a problem of resemblance with the initial mark). A phenomenology (or a hermeneutics) of this sign-impression is possible at the limit of what Husserl terms a hyletic discipline.
The third use of the mark: the corporeal, cerebral, cortical imprint, as discussed by neuroscience. For the phenomenology of the affection-impression, these corporeal imprints are the object of a presupposition concerning external causation, a presupposition whose status is extremely difficult to establish. We shall speak in this instance of a substratum, to indicate the connection of a particular sort between the impressions stemming from the world of experience and the material imprints in the brain belonging to the neurosciences.13 I shall say nothing more about this here but only point out the differences among these three uses of the indistinct idea of trace: trace written on a material support; affection-impression “in the soul”; corporeal, cerebral, cortical imprint. This is, to my mind, the ineluctable difficulty attached to the status of the “imprint in the soul” as in a block of wax. It is no longer possible today to avoid the problem of the relations between the cerebral imprint and the experienced impression, between the preservation-storage and the perseverance of the initial affection. I hope to show that this problem, inherited from the old debate over the relations between the soul and the body, and a problem audaciously assumed by Bergson in Matter and Memory, can be posed in terms other than those opposing materialism to spiritualism. Are we not dealing with two different readings of the body, of corporeality—the body as object confronting the body as lived—the parallel now shifting from the ontological plane to the linguistic or semantic plane?
Aristotle: “Memory Is of the Past”
Aristotle’s treatise Peri mnēmēs kai anamnēsēos, which has come down to us under the Latin title De memoria et reminiscentia, part of a collection of nine small treatises the tradition has named Parva naturalia, can be placed against the eristic and dialectic backdrop inherited from Plato.14 Why a double title? To distinguish, not the persistence of memories in relation to their recall, but their simple presence to mind (which I shall later call simple evocation in my phenomenological sketch) in relation to recollection as a search.
Memory, in this particular sense, is directly characterized as affection (pathos), which distinguishes it precisely from recollection.15
The first question raised is that of the “thing” remembered; it is here that the key phrase that will accompany my entire investigation is announced: “But memory is of the past” (449b15).16 It is the contrast with the future of conjecture and expectation and with the present of sensation (or perception) that imposes this major characterization. And it is under the authority of ordinary language (“No one would say. . . . Rather he says simply . . .”) that the distinction is made. Even more forcefully: it is “in one’s soul”17 that one has heard, or perceived, or thought this before (proteron) (449b23). This temporal mark raised to the level of language belongs to what below I shall call declarative memory. This mark is repeatedly stressed: just as it is true that we remember “without actually exercising (knowledge and perception)” (449b19), so it must also be emphasized that there is memory “when time has elapsed” (449b26).18 In this regard, humans share simple memory with certain animals, but all do not have the perception (aisthēsis) of time (449b29). This perception consists in the fact that the mark of anteriority implies the distinction between before and after, earlier and later. Now “earlier and later are in time [en chrono]” (450a21). This is in complete agreement with the analysis of time in Physics 4.11, according to which it is in perceiving movement that we perceive time; but time is perceived as other than movement only if we determine (horizomen) it (Physics 218b30),19 that is to say, if we distinguish two instants, one as earlier, the other as later.20 On this point, the analysis of time and the analysis of memory overlap. The second question concerns the relation between memory and imagination. They are tied together as a result of belonging to the same part of the soul, the sensible soul, following a method of division previously practiced by Plato.21 But the difficulty lies elsewhere: the proximity between the two problematics gives new strength to the old aporia concerning the mode of presence of the absent: “One might be puzzled how, when the affection is present but the thing is absent, what is not present is ever remembered” (450a25–26).
Aristotle responds to this aporia with what appears evident (dēlon) to him, namely, that the affection produced “by means of perception in the soul and in that part of the body which contains [it]” (450a26–27)22 should be considered a sort of picture (zōgraphema), “the having of which we say is memory” (450a30). Here, expressed in new terms that will interest us below, we find saddled up again the well-known problematic of the eikōn and, with it, the imprint (tupos), which is linked to the metaphor of the stamp and the seal. Nevertheless, in contrast to the Theaetetus, which places the imprint “in the souls”—even if it means treating them as impregnable—Aristotle connects the body to the soul and develops on this dual basis a rapid typology of the various effects of imprints (450b1–11). But our author has not finished with this metaphor. A new aporia arises: if this is the case, he asks, what is it that we remember? Is it the affection or the thing that produced it? If it is the affection, then it is not something absent one remembers; if it is the thing, then how, while perceiving the impression, could we remember the absent thing that we are not at present perceiving? In other words, while perceiving an image, how can we remember something distinct from it?
The solution to this aporia resides in the introduction of the category of otherness, inherited from the Platonic dialectic. The addition of the notion of imprint to that of drawing, of inscription (graphē)23 we would say today, sets us on the path toward a solution. It belongs to the notion of inscription that it contain a reference to the other; the other-than-affection as such. Absence, as the other of presence! Let us, Aristotle says, consider an example: a drawing of an animal. We can read this drawing in two ways: either we can consider it in itself, as a simple image drawn on a support, or as an eikōn (“a copy,” both of our translators write). We can do this because the inscription consists in both things at once: it is itself and the representation of something else (allou phantasma). Here, Aristotle’s vocabulary is precise: he reserves the term phantasma for the inscription itself and that of eikōn for the reference to the inscription’s other.24
The solution is clever but it has its own difficulties. The metaphor of the imprint, of which that of inscription is held to be a variant, invokes “movement” (kinēsis), from which the imprint results. This movement invokes an external cause (someone, something has made the imprint), while the double reading of the drawing, of the inscription, implies a division within the mental image, today we would say a double intentionality. This new difficulty seems to me to result from the competition between the two models, the impression and the inscription. The Theaetetus had paved the way for their confrontation by treating the imprint itself as a signifying mark, a sēmeion. It was then in the sēmeion itself that the external causation of the blow (kinēsis) and the actual meaning of the mark (sēmeion) were merged. The secret discordance between the two models emerges in Aristotle’s text when the production of the affection is set over against the iconic signification which both translators interpret as copy, hence as resemblance. This conjunction between (external) stimulation and (internal) resemblance will remain, for us, the crux of the entire problematic of memory.
The contrast between the two chapters of Aristotle’s treatise—mnēmē and anamnēsis—is more apparent than their belonging to one and the same problematic. The distinction between mnēmē and anamnēsis rests on two things: on the one hand, the simple memory arises in the manner of an affection, while recollection25 consists in an active search. On the other hand, the simple memory is under the dominion of the agent of the imprint, whereas movements and the entire sequence of changes that will be discussed have their principle in us. Nevertheless, the connection between the two chapters is secured by the role played by temporal distance: the act of remembering (mnēmoneuein) is produced when time has elapsed (prin khronisthēnai) (451a30). And it is this interval of time, between the initial impression and its return, that recollection traverses. In this sense, time indeed remains the factor common to memory as passion and to recollection as action. This factor, it is true, is almost lost from sight in the details of the analysis of recollection. The reason for this is the emphasis placed from here on, on the “how,” on the method of effective recollection.
In a general sense, “acts of recollection happen because one change [kinēsis] is of a nature to occur after another” (451b10).26 Now this succession can take place by necessity or out of habit; a certain margin of variation, a matter we shall return to below, is thereby preserved. Having said this, the priority given to the methodical side of the search (a term dear to all the Socratics) explains the insistence on the choice of a starting point for the course of the recollection. The initiative of the search thus stems from our “capacity for searching.” The starting point remains under the command of the explorer of the past, whether the connection that follows is the result of necessity or of habit. What is more, along this course, several paths remain open leading from this same starting point. The metaphor of making one’s way is thus induced by that of change. This is why the quest can mistakenly take the wrong track, and luck can always play a role. But the question of time is not lost from sight during these exercises of methodical remembering: “the main thing is that one must know the time” (452b7). This knowledge has to do with the measurement of the intervals elapsed, whether precise or indeterminate. In both cases, the estimation of more or less is part and parcel of this knowledge. And this estimation depends on the power to distinguish and compare magnitudes, whether of greater or smaller distances or dimensions. This estimation extends to the notion of proportion. Aristotle’s words confirm the thesis that the notion of temporal distance is inherent in the essence of memory and assures the distinction in principle between memory and imagination. Moreover, the role played by the estimation of lapses of time underscores the rational side of recollection: this “search” constitutes “a sort of reasoning [sullogismos]” (453a13–14). This does not prevent the body’s being involved in the sort of affection that is displayed in the hunt for the image (phantasma) (453a16).
In contrast to a reductive reading, this approach has produced a number of traditions of interpretation. First, the tradition of ars memoriae, which consists, as we shall say in chapter 2, in a form of memory training in which the operation of memorization prevails over the recollection of individual events of the past. Second, the associationism of modern philosophers, which, as Sorabji’s commentary underscores, draws solid support from Aristotle’s text. The text, however, also leaves room for a third conception, in which the accent is placed on the dynamism, the invention of connections, as in Bergson’s analysis of “the effort to recall.”
At the end of our reading and interpretation of Aristotle’s De memoria et reminiscentia, we are in a position to attempt an evaluation of the contribution of this treatise to a phenomenology of memory.
Its major contribution lies in the distinction between mnēmē and anamnēsis. We will encounter it below in a different vocabulary, with the terms “simple evocation” and “effort to recall.” By drawing a line in this way between the simple presence of memories and the act of recollection, Aristotle has preserved for all time a space for discussion worthy of the fundamental aporia brought to light by the Theaetetus, namely, the presence of the absent. The results of his contribution to this discussion are mixed. On the one hand, he sharpened the point of the enigma by making the reference to time the distinctive note of memory in the field of the imagination. With memory, the absent bears the temporal mark of the antecedent. On the other hand, by assuming the category of eikōn for the framework of the discussion, in connection with the category of tupos, he is in danger of pursuing the aporia to the point of impasse. The impasse is even twofold. On the one hand, throughout our investigation there will remain the troublesome question of determining whether the relation between the memory-image and the initial impression is one of resemblance, even of a copy. Plato had approached this difficulty by taking as his target the deceit inherent in this kind of relation, and in the Sophist he had even tried to distinguish between two mimetic arts: the fantastic art, deceitful by nature, and the eikastic art, capable of veracity. Aristotle appears to be unaware of the risks of error or illusion attaching to the conception of eikōn centered on resemblance. By holding at bay the misfortunes of the imagination and of memory, he may have wanted to shield these phenomena from the quarrels fomented by the sophists, reserving his reply and his attacks on them for his Metaphysics, principally in the framework of the problem of the self-identity of ousia. However, by not taking into account the degrees of fallibility belonging to memory, he removed the notion of iconic resemblance from the discussion. Another impasse: by taking for granted the tie between eikōn and tupos, he adds to the difficulties of the image as a copy those belonging to the notion of the imprint. But what of the relation between the external cause—“motion”—producing the imprint and the initial affection targeted by and in memory? To be sure, Aristotle has made great strides in the discussion by introducing the category of otherness into the very heart of the relation between the eikōn, reinterpreted as an inscription, and the initial affection. Having done this, he begins to advance the concept of resemblance, which, moreover, had not been challenged. But the paradoxes of the imprint will continue to reemerge, primarily with the question of the material causes of the anamnēsis of memory, prior to its recall.
As for anamnēsis, Aristotle has presented under this term the first analytical description of the mnemonic phenomenon of recollection, which is contrasted to the simple evocation of a memory that comes to mind. The richness and subtlety of his description place him at the head of a wide range of schools of thought seeking a model of interpretation for modes of connection arising from “necessity” or from “habit.” The associationism of the British empiricists is only one of these schools.
The astonishing thing, however, is that Aristotle retained the very term anamnēsis—one of the key words of Plato’s philosophy, from the Meno through the other great dialogues—to describe recollection as it operates under ordinary conditions. How are we to explain this faithful use of terms? Reverence due to his teacher? An appeal to authority suitable for covering an analysis that, nonetheless, naturalizes the grandiose vision of a knowledge lost at birth and recalled by study? Worse: betrayal disguised as faithfulness? Conjectures are endless. But none of those just mentioned goes beyond the level of the psychology of the author. Each draws its plausibility from the presumed thematic tie held to exist between the anamnēsis of Plato and that of Aristotle. This thematic tie is twofold: it is, first of all, on the plane of the aporia, the heritage of eikōn and of tupos, coming from the Theaetetus and the Sophist. For Plato, these categories were held to account for the possibility of sophistry and for the very existence of the Sophist, and so to stand in the position of counterpoint to the theory of reminiscence, which accounted only for the happy memory of the young slave of the Meno. With Aristotle, eikōn and tupos are the only categories available to account for the functioning of everyday memory; they no longer designate simply an aporia, but the direction in which this aporia has to be resolved. There is, however, an even stronger connection between Plato and Aristotle than that of the aporia on the path toward resolution. This tie has to do with their faithfulness to Socrates in the use of the two emblematic terms: “learning” and “seeking.” One must first have “learned” and then painfully “seek.” Because of Socrates, Aristotle was unable to, nor did he want to, “forget” Plato’s anamnēsis.
A PHENOMENOLOGICAL SKETCH OF MEMORY
Allow me to open the following sketch by making two remarks.
The first is in the guise of a warning against the tendency of many authors to approach memory on the basis of its deficiencies, even its dysfunctions, tendencies whose legitimate place we will indicate later.27 It is important, in my opinion, to approach the description of mnemonic phenomena from the standpoint of the capacities, of which they are the “happy” realization.28 In order to do this, I shall present in the least scholarly manner possible the phenomena that, in the ordinary language of everyday life, are placed under the heading of memory. What, in the final analysis, will justify taking this position in favor of “good” memory is my conviction, which the remainder of this study will seek to establish, that we have no other resource, concerning our reference to the past, except memory itself. To memory is tied an ambition, a claim—that of being faithful to the past. In this respect, the deficiencies stemming from forgetting, which we shall discuss in good time, should not be treated straight away as pathological forms, as dysfunctions, but as the shadowy underside of the bright region of memory, which binds us to what has passed before we remember it. If we can reproach memory with being unreliable, it is precisely because it is our one and only resource for signifying the past-character of what we declare we remember. No one would dream of addressing the same reproach to imagination, inasmuch as it has as its paradigm the unreal, the fictional, the possible, and other nonpositional features. The truthful ambition of memory has its own merits, which deserve to be recognized before any consideration is given to the pathological deficiencies and the nonpathological weaknesses of memory, some of which will be examined in the next section of this study, even before the confrontation with the deficiencies examined in the following study under the heading, abuses of memory. To put it bluntly, we have nothing better than memory to signify that something has taken place, has occurred, has happened before we declare that we remember it. False testimonies, which we shall discuss in the second part, can be unmasked only by a critical agency that can do nothing better than to oppose those accounts reputed to be more reliable to the testimony under suspicion. For, as will be shown, testimony constitutes the fundamental transitional structure between memory and history.
Second remark. Contrary to the polysemy, which, at first sight, seems sufficient to discourage even the most modest attempt at ordering the semantic field encompassed by the term “memory,” it is possible to sketch a splintered, but not radically dispersed, phenomenology in which the relation to time remains the ultimate and sole guideline. But this guideline can be held with a firm hand only if we succeed in showing that the relation to time of the various mnemonic modes encountered by our description is itself susceptible to a relatively well-ordered typology that is not exhausted, for example, by the case of the memory of a one-time event that occurred in the past. This second wager of our undertaking builds upon the minimal coherence of the assertion borrowed from Aristotle at the beginning of this study, according to which memory “is of the past.” But the being of the past can be said in many ways (in keeping with the famous passage from Aristotle’s Metaphysics that “being is said in many ways”).
The first expression of the splintered nature of this phenomenology stems from the object-oriented character of memory: we remember something. In this sense a distinction must be made in language between memory (la mémoire) as intention and memory (le souvenir) as the thing intended. We say memory (la mémoire) and memories (les souvenirs). Fundamentally, what is at issue here is a phenomenology of memories. In this regard, Latin and Greek use the preterite forms (genomenou, praeterita). It is in this sense that I speak of past “things.” Indeed once the past has been distinguished from the present in the memory of memories, then it is easy for reflection to distinguish at the heart of remembering the question “What?” from “How?” and from “Who?” following the rhythm of our three phenomenological chapters. In Husserlian terminology this is the distinction between the noesis of remembering and the noema of memories.
The first feature characterizing the domain of memories is their multiplicity and their varying degrees of distinctness. Memory in the singular is a capacity, an effectuation; memories are in the plural: we have memories (it is even said, unkindly, that the old have more memories than the young but less memory!). Later we shall evoke Augustine’s brilliant description of memories that spill over the threshold of memory, presenting themselves one by one or in bunches according to the complex relations of their themes or circumstances, or in sequences more or less amenable to being put into narrative form. In this regard, memories can be treated as discrete forms with more or less discernible borders, set off against what could be called a memorial backdrop, which can be a source of pleasant occupation in states of anamnēsis.
The most important feature, however, is the following: it has to do with the privilege spontaneously accorded to events among all the “things” we remember. In terms of the analysis we shall later borrow from Bergson, the “thing” remembered is plainly identified with a singular, unrepeatable event, for example a given reading of a memorized text. Is this always the case? To be sure, as we shall say in conclusion, the memory-event is in a way paradigmatic, to the extent that it is the phenomenal equivalent of a physical event. The event is simply what happens. It takes place. It passes and occurs (se passe). It happens, it comes about. It constitutes what is at stake in the third cosmological antinomy of the Kantian dialectic: either it results from something prior in accord with necessary causation or else it proceeds from freedom, in accord with spontaneous causation. On the phenomenological level, on which we have situated ourselves here, we say that we remember what we have done, experienced, or learned in a particular instance. But a range of typical cases unfolds between the two extremes of singular events and generalities, which can be termed “states of affairs.” Still closely resembling a unique event, we find discrete appearances (a certain sunset one particular summer evening), the singular faces of our loved ones, words heard according to their manner of utterance each time new, more or less memorable meetings (which we shall divide up again below following other criteria of variation). Things and people do not simply appear, they reappear as being the same, and it is in accordance with this sameness of reappearing that we remember them. In the same way, we recall names, addresses, and telephone numbers. Memorable meetings offer themselves to be remembered due less to their unrepeatable singularity than to their typical resemblance, even their emblematic character: a composite image of waking up in the morning in the house in Combray permeates the opening pages of Proust’s Remembrance of Things Past. Next comes the case of “things” we have learned and so acquired. In this way, we say that we still remember the table of Greek and Latin declensions and conjugations, or German and English irregular verbs. Not to have forgotten them is to be able to recite them without learning them all over again. These examples link up with the opposite pole, that of “states of affairs,” which, in the Platonic and Neoplatonic tradition to which Augustine still belongs, constitute the paradigmatic examples of Reminiscence. The canonical text for this tradition remains Plato’s Meno and the famous episode of the young slave’s rediscovery of certain noteworthy geometrical propositions. At this level, remembering and knowing completely coincide with one another. But states of affairs do not consist only in abstract generalities, in notions. Made the target of critique, as we shall say later, the events considered by documentary history display a propositional form that gives them the status of fact. It is then a matter of the “fact that . . .” things happened this way and not some other way. These facts can be said to be acquired; even, in the design of Thucydides, elevated to the rank of an “everlasting possession.” In this way, within the framework of historical knowledge, events tend to link up with “states of affairs.”
Given this diversity of past “things,” by what features are these “things”—these praeterita—recognized as being “of the past”? A new series of modes of dispersion characterize this “being of the past” common to all our memories. To guide our passage through the polysemic field of memory, I propose a series of oppositional pairs, constituting something like a rule-governed typology. This will obey an organizing principle capable of justification apart from its implementation, as is the case with Max Weber’s ideal types. If I were to seek terms of comparison, I would first think of analogy in Aristotle, halfway between simple homonymy, relegated to the dispersion of meaning, and polysemy, structured by a semantic core that would be identified by a genuine semiotic reduction. I would also think of Wittgenstein’s “family resemblance.” The reason for the relative indeterminacy of the epistemological status of the classification proposed has to do with the interconnection between preverbal experience—what I call lived experience, translating the Erlebnis of Husserlian phenomenology—and the work of language that ineluctably places phenomenology on the path of interpretation, hence of hermeneutics. Now the “working” concepts that prime the interpretation and direct the ordering of the “thematic” concepts proposed here, escape the mastery of meaning that a total reflection would want to command. More than others, the phenomena of memory, so closely connected to what we are, oppose the most obstinate of resistances to the hubris of total reflection.29
The first pair of oppositions is formed by habit and memory. It is illustrated in contemporary philosophy by the famous distinction between mémoire-habitude (memory as habit) and mémoire-souvenir (memory as distinct recollection) proposed by Bergson. We shall temporarily bracket the reasons why Bergson presents this opposition as a dichotomy. We shall instead follow the counsel of the experience least charged with metaphysical presupposition, for which habit and memory form two poles of a continuous range of mnemonic phenomena. What forms the unity of this spectrum is the common feature of the relation to time. In each of the opposing cases an experience acquired earlier is presupposed; however, in the case of habit what is acquired is incorporated into the living present, unmarked, unremarked as past. In the other case, a reference is made to the anteriority of the prior acquisition. In both cases, then, it remains true that memory “is of the past,” but according to two distinct modes—unmarked and marked—of reference to the place in time of the initial experience.
If I place the pair habit/memory at the start of my phenomenological sketch, this is because it provides the first opportunity to apply to the problem of memory what, since the introduction, I have called the conquest of temporal distance, a conquest relying on a criterion that can be described as a gradient of distantiation. The descriptive operation then consists in arranging experiences relative to temporal depth, beginning with those in which the past adheres, so to speak, to the present and continuing on to those in which the past is recognized in its pastness as over and done with. Let me refer, as so many others have done, to the famous pages in chapter 2 of Matter and Memory devoted to the distinction between “two forms of memory.”30 Like Augustine and the ancient rhetoricians, Bergson places himself in the situation of reciting a lesson learned by heart. Habit-memory is then the one we employ when we recite the lesson without evoking one by one each of the successive readings of the period of learning. In this case, the lesson learned “is part of my present, exactly like my habit of walking or of writing; it is lived and acted, rather than represented” (91). On the other hand, the memory of a particular reading, of a given phase of memorization, presents “none of the marks of a habit”: “It is like an event in my life; its essence is to bear a date, and consequently to be unable to occur again” (90). “The image, regarded in itself, was necessarily at the outset what it always will be” (90). And again: “Spontaneous recollection is perfect from the outset; time can add nothing to its image without disfiguring it; it retains its memory in place and date” (95). In short: “The memory of a given reading is a representation, and only a representation” (91); whereas the lesson learned is, as just said, “acted” rather than represented, it is the privilege of representation-memory to allow us “in the search for a particular image [to] remount the slope of our past” (92). To memory that repeats is opposed memory that imagines: “To call up the past in the form of an image, we must be able to withdraw ourselves from the action of the moment, we must have the power to value the useless, we must have the will to dream. Man alone is capable of such an effort” (94).
This is a text of great richness. In its crystalline sobriety, it posits the more extensive problem of the relation between action and representation, of which the exercise of memorization is only one aspect, as I will state in the next chapter. In doing this, Bergson underscores the kinship between the lesson learned by heart and “my habit of walking or of writing.” What is stressed in this way is the set to which recitation belongs, that of knowing-how, which includes in an array of different modes the common feature of being ready to . . . , without having to repeat the effort of learning again, of re-learning; as such, these modes are able to be mobilized in a range of different occasions, just as they are open to a degree of variability. It is to these instances of knowing-how that, among the vast panoply of uses of the word “memory,” we apply one of its accepted senses. In this way, the phenomenologist will be able to distinguish “remembering how . . .” and “remembering that . . .” (an expression that will lend itself to further distinctions). This vast empire covers forms of know-how on very different levels: we encounter first corporeal capacities and all the modalities of “I can” which are considered in my own phenomenology of the “capable human being”: being able to speak, being able to intervene in the course of affairs, being able to recount, being able to ascribe an action to oneself by making oneself its actual author. To this must be added social customs, mores, all the habitus of life in common, part of which is involved in the social rituals belonging to phenomena of commemoration, which we will later contrast to the phenomena of rememoration, assigned to private memory alone. Several polarities intersect in this way. We will encounter others equally significant in the framework of the present consideration, where the accent falls on the application of the criterion of temporal distantiation.
The fact that, on the phenomenological plane, we are considering a polarity and not a dichotomy is confirmed by the eminent role held by phenomena situated between the two poles that Bergson opposes following his customary method of division.
The second set of opposites is constituted by the pair evocation/search. By evocation let us understand the unexpected appearance of a memory. Aristotle reserved for this the term mnēmē, reserving anamnēsis for what we shall later call search or recall. And he defined mnēmē as a pathos, as an affection: it happens that we remember this or that, on such and such an occasion; we then experience a memory. Evocation is an affection, therefore, in contrast to the search. In other words, abstracting from this polarity, evocation as such bears the weight of the enigma that set in motion the investigations of Plato and Aristotle, namely, the presence now of the absent that was earlier perceived, experienced, learned. This enigma must be provisionally disassociated from the question raised by the perseverance of the first affection, illustrated by the famous metaphor of the imprint of the seal and, consequently, from the question of whether the faithfulness of a memory consists in the resemblance of the eikōn to the first imprint. Neuroscience has taken up this problem under the title of mnestic traces. This problem must not monopolize our attention: phenomenologically speaking, we know nothing of the corporeal, and more precisely cortical, substratum of evocation, nor are we clear about the epistemological status of the correlation between the formation, conservation, and activation of these mnestic traces and the phenomena that fall under the phenomenological gaze. This problem belonging to the category of material causation should be bracketed as long as possible. I shall wait until the third part of this work before confronting it. However, what must be brought to the fore, following Aristotle, is the reference to the anteriority of the “thing” remembered in relation to its present evocation. The cognitive dimension of memory, its character of knowing, lies in this reference. It is by virtue of this feature that memory can be held to be trustworthy or not and that properly cognitive deficiencies are to be accounted for, without our rushing to construe them according to a pathological model, under the heading of this or that form of amnesia.
Let us move to the other pole of the pair evocation/search. This is what was designated by the Greek term anamnēsis. Plato had turned it into myth by tying it to a prenatal knowledge from which we are said to have been separated by a forgetting that occurs when the life of the soul is infused into a body—described, moreover, as a tomb (soma-sēma)—a forgetting from birth, which is held to make the search a relearning of what has been forgotten. In the second chapter of the treatise analyzed above, Aristotle naturalizes anamnēsis, so to speak, bringing it closer to what in everyday experience we term recollection. Along with all the Socratics, I designate recollection by means of the enigmatic term of searching (zētēsis). The break with Platonic anamnēsis is nevertheless not complete, to the extent that the ana of anamnēsis signifies returning to, retaking, recovering what had earlier been seen, experienced, or learned, hence signifies, in a sense, repetition. Forgetting is thus designated obliquely as that against which the operation of recollecting is directed. The work of anamnēsis moves against the current of the river Lēthē. One searches for what one fears having forgotten temporarily or for good, without being able to decide, on the basis of the everyday experience of recollection, between two hypotheses concerning the origin of forgetting. Is it a definitive erasing of the traces of what was learned earlier, or is it a temporary obstacle—eventually surmountable—preventing their reawakening? This uncertainty regarding the essential nature of forgetting gives the search its unsettling character.31 Searching is not necessarily finding. The effort to recall can succeed or fail. Successful recollection is one of the figures of what we term “happy” memory.
With regard to the mechanism of recollection, I mentioned, within the framework of my commentary on Aristotle’s treatise, the range of procedures employed, from quasi-mechanical association to the work of reconstruction which Aristotle compares to sullogismos, to argumentation.
I would like to give a modern echo here to the ancient texts. Once again I shall refer to Bergson, reserving for later a thorough examination of the fundamental theory of Matter and Memory, which will encompass the borrowings made here from Bergson’s analyses. I am thinking in this regard of the essay titled “Intellectual Effort” in Mind-Energy,32 principally those pages devoted to “the effort of memory.”
The primary distinction is between laborious recollection and spontaneous recollection (188–203), where spontaneous recollection can be considered the zero-degree of searching and laborious recollection its purposeful form. The major interest of Bergson’s essay lies in the struggle against the reduction performed by associationism of all the forms of searching to the most mechanical among these. The distinction between the two forms of recollection is set within a more extensive inquiry, placed under a single question: “What is the intellectual characteristic of intellectual effort?” (187). Whence the title of the essay. The scope and the precision of the question deserve to be underscored in turn. On the one hand, the recollection of a memory belongs to a vast family of mental facts: “When we call to mind past deeds, interpret present actions, understand a discourse, follow someone’s train of thought, attend to our own thinking, whenever, in fact, our mind is occupied with a complex system of ideas, we feel we can take up two different attitudes, one of tension, the other of relaxation, and they are mainly distinguished by the feeling of effort which is present in the one and absent from the other” (186). On the other hand, the precise question is this: “Is the play of ideas the same in each case? Are the intellectual elements of the same kind, and have they the same relations among themselves?” (186). The question, we see, cannot fail to interest contemporary cognitive science.
If the question of recollection comes first in the study applied to the various types of intellectual labor, this is because the gradation “starting with the easiest, which is reproduction, and ending up with the most difficult, which is production and invention” (188) is most marked here. What is more, the essay can use as a basis the distinction made in Matter and Memory between “a series of different ‘planes of consciousness,’ beginning with the plane of ‘pure memory’ not yet translated into distinct images, and going down to the plane where the same memory is actualized in nascent sensations and incipient movements” (188). The voluntary evocation of a memory consists precisely in this traversal of planes of consciousness. A model is then proposed for distinguishing the role of automatic, mechanical recall from that of reflection, of intelligent reconstruction, intimately mingled in ordinary experience. It is true that the example chosen is recalling a text learned by heart. It is, therefore, at the time of learning that the split occurs between the two types of reading; in the analytical reading, there is a hierarchy between the dominant idea and the subordinate ideas, to which Bergson relates the famous concept of a dynamic scheme: “I mean by this, that the idea does not contain the images themselves so much as the indication of what we must do to reconstruct them” (196). Exemplary in this regard is the chess-player, who can play several games at once without looking at the board: “What is present to the mind of the player is a composition of forces, or rather a relation between allied or hostile forces” (198). Each game is thus memorized as a whole following its own profile. It is, therefore, in the method of learning that we must seek the key to the phenomenon of recollection, for example, that of the troublesome search for a recalcitrant name: “an impression of strangeness, but not of strangeness in general” (199). The dynamic scheme acts as a guide “indicating a certain direction of effort” (200). In this example, as in many others, “the effort of memory appears to have as its essence the evolving of a scheme, if not simple at least concentrated, into an image with distinct elements more or less independent of one another” (201). Such is the manner of traversing the planes of consciousness, “a descent of the scheme towards the image” (202). We can then say that “the effort of recall consists in converting a schematic idea, whose elements interpenetrate, into an imaged idea, the parts of which are juxtaposed” (203). It is in this that the effort of recall constitutes a case of intellectual effort and is associated with the effort of intellection examined in chapter 2 of Matter and Memory: “Whether we are following an argument, reading a book or listening to a discourse” (205), the “feeling of effort, in intellection, is produced on the passage from the scheme to the image” (211). What remains to be examined is what makes the work of memory, intellection, or invention an effort, namely, the difficulty signaled by the discomfort experienced or the obstacle encountered, finally the properly temporal aspect of slowing down or of delay. Longstanding combinations resist the reworking required by the dynamic scheme, as do the images themselves in which the schema seeks to be inscribed. Habit resists invention: “In this peculiar kind of hesitation is likely to be found intellectual effort” (215). And “we may conceive that this indecision of the mind is continued in a disquietude of the body” (222). Arduousness thus has its own affectively experienced temporal mark. There is pathos in zētēsis, “affection” in “searching.” In this way, the intellectual and the affective dimensions of the effort to recall intersect with one another, as they do in every other form of intellectual effort.
At the end of this study of recollection, I would like to make a brief allusion to the relation between the effort to recall and forgetting (before we have the opportunity in the third part of this work to engage in a proper discussion of the problems concerning forgetting, problems we encounter here in random order).
It is, in fact, the effort to recall that offers the major opportunity to “remember forgetting,” to anticipate the words of Augustine. Searching for a memory indeed attests to one of the major finalities of the act of remembering, namely, struggling against forgetting, wresting a few scraps of memory from the “rapacity” of time (Augustine dixit), from “sinking” into oblivion (oubli). It is not only the arduousness of the effort of memory that confers this unsettling character upon the relation, but the fear of having forgotten, of continuing to forget, of forgetting tomorrow to fulfill some task or other; for tomorrow, one must not forget . . . to remember. In the next chapter, what I will call the duty of memory consists essentially in a duty not to forget. In this way, a good share of the search for the past is placed under the sign of the task not to forget. More generally, the obsession of forgetting, past, present, and future, accompanies the light of happy memory with the shadow cast by an unhappy memory. For meditating memory—Gedächtnis—forgetting remains both a paradox and an enigma. A paradox, as it is unfolded by Augustine the rhetorician: how can we speak of forgetting except in terms of the memory of forgetting, as this is authorized and sanctioned by the return and the recognition of the “thing” forgotten? Otherwise, we would not know that we have forgotten. An enigma, because we do not know, in a phenomenological sense, whether forgetting is only an impediment to evoking and recovering the “lost time,” or whether it results from the unavoidable wearing away “by” time of the traces left in us by past events in the form of original affections. To solve the enigma, we would have not only to uncover and to free the absolute ground of forgetting against which the memories “saved from oblivion” stand out, but also to articulate this non-knowledge concerning the absolute ground of forgetting on the basis of external knowledge—in particular, that of the neurological and cognitive sciences—of mnestic traces. We shall not fail to return, at the appropriate time, to this difficult correlation between phenomenological knowledge and scientific knowledge.33
A separate and prominent place must be given to the distinction introduced by Husserl in The Phenomenology of the Consciousness of Internal Time between retention or primary memory and reproduction or secondary memory.34 We read of this distinction in the second section of the 1905 “Lectures on the Phenomenology of the Consciousness of Internal Time,” which form the first part of the work, supplemented by additions and complements from the years 1905–10. I have separated out those analyses that concern the object-side of memory, as the translation of Erinnerung by “memory” (souvenir) confirms, and, in the remainder of the present chapter, added to them Husserl’s reflections on the relation between memory and image. By separating this section out from the main context of the 1905 lectures, I remove it from the province of the subjective idealism that is grafted onto the reflexive side of memory (which I will examine later in the concluding chapter of this phenomenology of memory). I confess that this liberation cuts against the grain of the overall dynamic of the 1905 lectures, which, from the first to the third section, traverse a series of “levels of constitution” (Husserl, §34), gradually erasing the objective character of the constitution to the benefit of the self-constitution of the flow of consciousness. The “temporal objects”—in other words, the things that endure—then appear as “constituted unities” (Husserl, §37) in the pure reflexivity of the consciousness of internal time. My argument here is that the famous epoché with which the work opens and that results in bracketing objective time—the time that cosmology, psychology, and the other human sciences take as a reality, formal to be sure, yet of a piece with the realist status of the phenomena it frames—does not begin by laying bare a pure flow, but rather a temporal experience (Erfahrung) that has an object-oriented side in memory. The constitution at the first level is that of a thing that endures, however minimal this objectivity may be, first following the model of a sound that continues to resonate, then of a melody that one remembers after the fact. However, in each case, “something” endures. The epoché, to be sure, does expose pure experiences, “experiences of time” (Husserl, §2, 10). But in these experiences, “data ‘in objective time’ are meant” (ibid.). They are termed “objectivity” (ibid.) and contain “a priori truths that pertain to the different constitutive moments of the objectivity” (ibid.). If from the start of our reading, the reference to this “objective” aspect appears provisional, this is because a radical question is raised, that of the “origin of time” (11), which is intended to be kept out of the realm of psychology, without thereby slipping into the orbit of Kantian transcendentalism. The question posed by the experience of a sound that continues and of a melody that returns concerns the sort of persistence by which “what we perceive remains present to us for a time, but not without undergoing modification” (Husserl, §3, 11). The question is: what is it for something that endures to remain? What is temporal duration? This question is no different than those posed by William James and Henri Bergson in similar terms: endure, persist, remain. What modification is this? Is it a sort of association (Brentano)? Is it a sort of recapitulative comparison with the last sound (W. Stern)? These solutions can be discarded but not the problem, namely, “the apprehension of transcendent temporal objects that are extended over a duration” (§7, 23). Let us call these objects “temporal objects” (Zeitobjekten) on the basis of which the question of the constitution of time will later be posed, when it will be considered to be a duration undifferentiated by the objects that endure. The analysis will then shift from the perception of the duration of something to a study of the duration of perception as such. It will then no longer be the sound, the melody that will be thematized but rather their unobjectifiable duration. Just before this change of emphasis, the noteworthy distinction between immediate memory or retention and secondary memory (recollection) or reproduction will become meaningful.
The experience described has a pivotal point, the present, the present of the sound that resonates now: “When it begins to sound, I hear it as now; but while it continues to sound it has an ever new now, and the now that immediately precedes it changes into a past” (§7, 25). It is this modification that constitutes the theme of the description. There is an “ever new” now. The situation described is in this regard no different from that considered by Augustine in book 11 of the Confessions: the modification is of the present. Of course, Augustine is unaware of the bracketing of every transcendent thesis and the reduction of the sound to a pure “hyletic datum” (§8, 25). But the idea that something begins and ceases, begins and “recedes” after it ends into the most distant past, is common. What is then proposed is the idea of “retention”: “In this sinking back, I still ‘hold onto it,’ have it in a ‘retention.’ And as long as the retention lasts, the tone has its own temporality” (25). At this stage of the analysis the two propositions coincide: the sound is the same, its duration is the same. Later, the second one will assimilate the first one. We will then pass from the phenomenology of memory to that of the consciousness of internal time. The transition is prepared by the remark that “I can direct my attention to the way in which it is given” (25). Then the “modes” and their continuity, in a “continual flow,” will move to the forefront. This, however, will not eliminate the reference to the now that, at the start of the analysis with which we are concerned here, is the phase of a sound, that phase termed “consciousness of the commencing tone” (ibid.): “The tone is given; that is, I am conscious of it as now” (25–26). At a later stage of the analysis, this stubborn reference to the present will attest to the reign of what Heidegger and those influenced by him denounce as a “metaphysics of presence.”35 On the level to which I am confining the analysis here, the reference to the present links up with the everyday experience that we have of things that begin, continue, and cease to appear. Beginning constitutes an undeniable experience. Without it, we could not understand the meaning of continuing, enduring, remaining, stopping. And always, there is something that begins and ceases. Moreover, the present is not to be identified with presence—in any metaphysical sense. The phenomenology of perception does not even have any exclusive right regarding the description of the present. The present is also the present of enjoyment and suffering and, more significantly for an investigation of historical knowledge, the present of initiative. The reproach that can legitimately be made to Husserl, at this preliminary stage of his analysis, is to have enclosed the phenomenology of the present within perceived objectivity at the expense of affective and practical objectivity. Within these limits, his thesis is simply that perception is not instantaneous, that retention is not a form of imagination, but consists in a modification of perception. The perception of something has a duration. The distance “from the actually present now-point” (§9, 27) is still a phenomenon of perception and not of imagination. It is with regard to something that we say it endures: “The ‘consciousness,’ the ‘experience,’ is related to its object by means of an appearance in which precisely the ‘object in its way of appearing’ stands before us” (§9, 28). The phenomenology of memory is initially that of memories, if by this is understood “the object in its way of appearing.” What is called present, past, are its “running-off characters” (§10, 29), eminently immanent phenomena (in the sense of a transcendence reduced to its hyletic status).
If a tension is observable in the analysis, before the appearance of the distinction between retention and remembering, it is between fixing on the actual now and the indivisibility into fragments of the phenomenon of running-off. But Husserl should not be reproached for this tension, as though it were the inconsequential result of a metaphysical complacency: it is constitutive of the phenomena described. We can indeed pass without stopping, like time itself, from one phase to the other of the duration of the same object, or stop at one phase: the beginning is simply the most remarkable of these stopping points; but cessation is just as remarkable. In this way, we begin doing something and we stop doing it. Acting, in particular, has its knots and its swells, its fits and its starts; acting is muscular. And in the smoother succession of perception, the distinction between beginning, continuing, and stopping is perfectly reasonable. It is as a beginning that the present makes sense and that duration amounts to a modification: “Since a new now is always entering on the scene, the now changes into a past; and as it does so the whole running-off continuity of pasts belonging to the preceding point moves ‘downwards’ uniformly into the depths of the past” (§10, 30). Is the term “source-point” used here (§11, 30)? This is within the framework of the relation beginning-continuing-ceasing. The impression is primal, in a nonmetaphysicial sense, in the sense of something that simply begins and by reason of which there is a before and an after. The present is continually changing, but it is also continually arising: what we call happening. On this basis, running-off is only “a retention of retention” (§11, 31). But the distinction beginning/continuing never ceases to signify, so that “this continuity itself is again an actually present point that is retentionally adumbrated,” which Husserl likens to the tail of a comet. We then speak of a duration that “is finished” (31). This end-point can indeed be analyzed in terms of a continuity of retentions; but as an end, it presents itself as a “now-apprehension,” as “the head attached to the comet’s tail” (32).36
What then of the eventual end of the attenuation that would be its disappearance? In evoking this, Husserl speaks of imperceptibility (32), thereby suggesting the limited character of the temporal field as a field of visibility. This remark is also valid for the diagram in §10: “No ending of retention is foreseen there” (Husserl’s note, 32), which, according to certain authors, would allow for both an admission that forgetting is unavoidable and that there is an unconscious persistence of the past.
In summary, to term “primal” the past instant proper to retention is to deny that this is a figuration in terms of images. It is this distinction that we will take up anew on the basis of the unpublished texts and in relation to a different cycle of analyses tied to the positional/nonpositional opposition. In the 1905 lectures the opposition between impressional and retentional predominates. This distinction suffices to separate the now of consciousness from the “just past” that gives a temporal extension to perception. An opposition to the imaginary is nevertheless already in place: in truth, it existed as early as the critique of Brentano in the first section. As for the distinction between impression and retention, the focus of our discussion here, it derives, according to Husserl, from an eidetic necessity. This is not given de facto: “we teach the a priori necessity that a corresponding perception, or a corresponding primal impression, precede the retention” (§13, 35). In other words, for something that endures, continuing presupposes beginning. One might raise certain “Bergsonian” reservations about the equivalence between the now and the point, but not about the distinction between beginning and continuing. This distinction is constitutive of the phenomenology of memory—of that memory of which it is said, “givenness of the past is memory” (§13, 36). And this givenness necessarily includes a moment of negativity: the retention is not the impression; the continuity is not a beginning. In this sense, retention is “not now”: “‘Past’ and ‘now’ exclude one another” (36). To endure is in a certain way to go beyond this exclusion. To endure is to remain the same. This is what is signified by the word “modification.”
It is in relation to this exclusion—to this primordial not-now—of the past nevertheless retained that a new kind of polarity is suggested within the not-now of memory itself. This is the polarity of primary memory and secondary memory, of retention and reproduction.
Reproduction assumes that the primary memory of a temporal object such as melody has “disappeared” and that it comes back. Retention still hangs onto the perception of the moment. Secondary memory is no longer presentation at all; it is re-presentation. It is the same melody but heard “as it were” (§14, 37). The melody heard earlier “in person” is now remembered, re-presented. The memory itself can in turn be retained in the mode of having just been remembered, re-presented, re-produced. All the distinctions suggested elsewhere between spontaneous and laborious evocation as well as those concerning degrees of clarity can be applied to this modality of secondary memory. The essential thing is that the reproduced temporal object has no longer a foot, so to speak, in perception. It has removed itself. It is really past. And yet it links up with, it follows after the present and its comet’s tail. The interval is what we name a lapse of time. At the time of the 1905 lectures and the 1905–10 supplements, reproduction is classified among the modes of imagination (Appendix II, 107–9). The distinction remains to be made between thematizing and de-thematizing imagination, the sole tie between them being absence, a major bifurcation recognized by Plato, in terms of mimetic art, in the distinction between the fantastic and the iconic. Speaking here of the “reproduction” of duration, Husserl implicitly evokes the differential thetic character of memory.37 The fact that reproduction is also imagination, is Brentano’s limited truth (§19): in negative terms, to reproduce is not to give in person. To be given once again is not to have just been given. The difference is no longer continuous, but discontinuous. The formidable question is then posed, that of knowing under what conditions “reproduction” is reproduction of the past. The difference between imagination and recollection depends on the answer to this question. It is then the positional dimension of recollection that makes the difference: “Recollection, on the other hand, posits what is reproduced and in this positing gives it a position in relation to the actually present now and to the sphere of the original temporal field to which the recollection itself belongs” (§23, 53). Husserl refers here to Appendix III: “The Nexus-Intentions of Perception and Memory—The Modes of Time-Consciousness.” At this price, the reproduced now can be said to “coincide” with a past now. This “double intentionality” corresponds to what Bergson and others have called recognition—the conclusion to a happy quest.
At this point, a meticulous analysis devoted to the distinction between Erinnerung and Vorstellung, collected in volume 23 of Husserliana, picks up from that of the second section of The Phenomenology of the Consciousness of Internal Time. I will return to it in the final section of this chapter in the context of the confrontation between memories and images.
I would like to complete this review of the polarities by considering one pair of opposed yet complementary terms, the importance of which will be fully revealed at the time of the transition from memory to history.
I am speaking of the polarity between reflexivity and worldliness. One does not simply remember oneself, seeing, experiencing, learning; rather one recalls the situations in the world in which one has seen, experienced, learned. These situations imply one’s own body and the bodies of others, lived space, and, finally, the horizon of the world and worlds, within which something has occurred. Reflexivity and worldliness are indeed related as opposite poles, to the extent that reflexivity is an undeniable feature of memory in its declarative phase: someone says “in his heart” that he formerly saw, experienced, learned. In this regard, nothing should be stripped from the assertion that memory belongs to the sphere of interiority—to the cycle of inwardness, to borrow Charles Taylor’s vocabulary in Sources of the Self.38 Nothing should be removed except the interpretive surplus of subjectivist idealism that prevents this moment of reflexivity from entering into a dialectical relation with the pole of worldliness. To my mind, it is this “presupposition” that burdens the Husserlian phenomenology of time, despite its ambition to be constituted without presuppositions, listening only to the teaching of the “things themselves.” This is a questionable effect of the epoché, which, under the guise of objectification, strikes worldliness. Actually, in defense of Husserl, it must be said that the phenomenology of the Lebenswelt, developed in Husserl’s last great book, partially eliminates the equivocation by restoring its primordial character to what we globally term the situation in the world, without, however, breaking with the transcendental idealism that marks the works of the middle period, culminating in Ideen I but already foreshadowed in The Phenomenology of the Consciousness of Internal Time.
The considerations which follow owe an immense debt to Edward Casey’s magisterial work, Remembering.39 The sole point of divergence separating me from Casey concerns the interpretation he draws from the phenomena he so marvelously describes: he thinks he must step outside the region permeated by the theme of intentionality and, along with it, by Husserlian phenomenology, under the sway of the existential ontology inaugurated by Heidegger in Sein und Zeit. Whence the opposition that guides his description of mnemonic phenomena, separating them into two great masses signaled by the titles “Keeping Memory in Mind” and “Pursuing Memory beyond Mind.” But what does “mind” (an English term so difficult to translate into French) signify? Does not this term refer to the idealist interpretation of phenomenology and to its major theme, intentionality? As a matter of fact, Casey accounts for the complementarity between these two great ensembles by inserting between them what he calls “mnemonic Modes,” namely, Reminding, Reminiscing, Recognizing. What is more, he makes no bones about calling his great work A Phenomenological Study. Allow me to add a word to confirm my profound agreement with Casey’s undertaking: above all, I admire the general orientation of the work, aimed at protecting memory itself from forgetfulness (whence the title of the introduction “Remembering Forgotten: The Amnesia of Anamnesis”—to which part four, “Remembering Re-membered,” provides a response). In this regard, the book is a plea for what I call “happy” memory, in contrast to descriptions motivated by suspicion or by the excessive primacy accorded to phenomena of deficiency, even to the pathology of memory.
I have nothing really new to say here concerning the reflexive pole of the pair considered here, to the extent that this title encompasses phenomena that have already appeared in the other pairs of opposites. One would have to trace them back to the polarity between one’s own memory and the collective memory of our next study. Moreover, it is with the latter, under the title of “Commemoration,” that Casey completes his “pursuit” of memory “beyond mind.” One would then have to collect under the heading of reflexivity the “right hand” term of each of the preceding pairs. In this way, in the opposition between habit and memory, the habitual side is less marked with regard to reflexivity: one exercises know-how without noticing it, without paying attention to it, without being mindful of it. When a performance is flubbed, then one is called to attention: mind your step! As for the pair evocation/recollection, reflexivity is at its height in the effort to recall; it is underscored by the feeling of arduousness tied to the effort. Simple evocation can, in this regard, be considered neutral or unmarked, inasmuch as the memory is said to arise as the presence of the absent. It can be said to be marked negatively in the case of spontaneous, involuntary evocation, well known to the readers of Proust’s Remembrance of Things Past, and even more so in the case of the obsessional irruptions considered in the next chapter. Evocation is no longer simply experienced (pathos) but suffered. “Repetition” in the Freudian sense is then the inverse of remembering, which can perhaps be compared, as the work of memory, to the effort of recollection described above.
The three “mnemonic modes” that Casey interposes between the intentional analysis of memory held captive, as he says, “in Mind” and the pursuit of memory “beyond Mind” constitute, in fact, transitional phenomena in memory, between the pole of reflexivity and the pole of worldliness.
What does the word reminding convey? There is no appropriate term in French, if not one of the uses of the word rappeler: this reminds me (me rappelle) of that, makes me think of that. Might we say memento, memory-aid, pense-bête, or in the experimental sciences, points of reference, reminders? Indeed, it stands for clues that guard against forgetting. They are distributed on either side of the dividing line between the inner and the outer; they are found, first, on the side of recollection, either in the frozen form of the more or less mechanical association by which one thing is recalled by means of another associated with it through a learning process, or as one of the “living” relays of the work of recollection. They are found a second time in the form of external points of reference for recall: photographs, postcards, diaries, receipts, mementos (the famous knot in the handkerchief!). In this way, these signposts guard against forgetting in the future: by reminding us what is to be done, they admonish us not to forget to do it (feed the cat!).
As for reminiscing, this is a phenomenon more strongly marked by activity than reminding; it consists in making the past live again by evoking it together with others, each helping the other to remember shared events or knowledge, the memories of one person serving as a reminder for the memories of the other. This memorial process can, of course, be internalized in the form of meditative memory, an expression that best translates the German Gedächtnis, with the help of a diary, memoirs or anti-memoirs, autobiographies, in which the support of writing provides materiality to the traces preserved, reanimated, and further enriched with unpublished materials. In this way, provisions of memories are stored up for days to come, for the time devoted to memories . . . The canonical form of reminiscing, however, is conversation in the province of the spoken word: “Say, do you remember . . . , when . . . you . . . we . . . ?” The mode of reminiscing thus unfolds along the same line of discursivity as simple evocation in its declarative stage.
There remains the third mnemonic mode, which Casey terms one of transition: recognizing. Recognizing appears at first as an important complement to recollection, its sanction one might say. We recognize as being the same the present memory and the first impression intended as other.40 In this way, we are referred back by the phenomenon of recognition to the enigma of memory as presence of the absent encountered previously. And the “thing” recognized is doubly other: as absent (other than presence) and as earlier (other than the present). And it is as other, emanating from a past as other that it is recognized as being the same as. This complex otherness itself presents degrees corresponding to the degrees of differentiation and distantiation of the past in relation to the present. The otherness is close to zero in the feeling of familiarity: one finds one’s bearings, one feels at ease, at home (heimlich) in the enjoyment of the past revived. The otherness is, in contrast, at its height in the feeling of strangeness (the famous Unheimlichkeit of Freud’s essay, the “uncanny”). It is maintained at its median degree when the event recalled is, as Casey says, traced “back there where it was” (125). This median degree announces, on the plane of the phenomenology of memory, the critical operation by which historical knowledge restores its object to the kingdom of the expired past, making of it what Michel de Certeau called “the absent of history.”
The small miracle of recognition, however, is to coat with presence the otherness of that which is over and gone. In this, memory is re-presentation, in the twofold sense of re-: turning back, anew. This small miracle is at the same time a large snare for phenomenological analysis, to the extent that this representation threatens to shut reflection up once again within the invisible enclosure of representation, locking it within our head, in the mind.
Nor is this all: the fact also remains that the recognized past tends to pass itself off as a perceived past, whence the strange fate of recognition to be able to be treated within the framework of the phenomenology of memory and within the framework of perception. There is no forgetting Kant’s famous description of the threefold subjective synthesis: apprehension, reproduction, recognition. Thus recognition assures the cohesion of the perceived itself. It is in similar terms that Bergson speaks of the unfolding of the dynamic scheme in images as a return to perception. We will come back to this in the third section of this chapter when we consider memories in the form of images.
Once we have run through the “mnemonic modes” that Casey’s typology places half-way between the phenomena that the phenomenology of intentionality (overburdened, in my opinion, by subjective idealism) is held to situate in Mind and those it seeks beyond Mind, we are faced with a series of mnemonic phenomena implying the body, space, the horizon of the world or of a world.
In my opinion, these phenomena do not take us out of the sphere of intentionality but reveal its nonreflexive dimension. I remember having experienced pleasure and pain in my body at one time or another in my past life; I remember having lived for a long time in a certain house in a certain town, to have traveled in a certain part of the world, and it is from here that I evoke all those elsewheres. I remember the expanse of a certain seascape that gave me the feeling of the vastness of the world. And, during a visit to an archeological site, I evoked the cultural world gone by to which these ruins sadly referred. Like the witness in a police investigation, I can say of these places, “I was there.”
Beginning with corporeal memory, let us recognize that it too is capable of being divided along the first axis of oppositions: from the body-as-habit to the body-as-event, so to speak. The present polarity of reflexivity/worldliness partially coincides with the former one. Corporeal memory can be “enacted” in the same manner as all the other modalities of habit, such as driving a car when I am at the wheel. It is modulated in accordance with all the variations of feelings of familiarity or of strangeness. But the ordeals, illnesses, wounds, and traumas of the past invite corporeal memory to target precise instances that call in particular upon secondary memory, upon recollection, and invite a recounting. In this regard, happy memories, especially erotic ones, leave no less a mark of their singular place in the elapsed past, without forgetting the promise of repetition that they contain. Corporeal memory is thus peopled with memories affected with varying degrees of temporal distantiation: the magnitude of the interval of time elapsed can itself be perceived, felt, in the mode of regret, of nostalgia. The moment of awakening, so magnificently described by Proust at the beginning of Remembrance of Things Past, is especially favorable for returning things and beings to the place assigned to them in space and in time the previous evening. The moment of recollection is then the moment of recognition. The latter, in its turn, can span all the degrees from tacit remembering to declarative memory, ready for narration once again.
The transition from corporeal memory to the memory of places is assured by acts as important as orienting oneself, moving from place to place, and above all inhabiting. It is on the surface of the habitable earth that we remember having traveled and visited memorable sites. In this way, the “things” remembered are intrinsically associated with places. And it is not by chance that we say of what has occurred that it took place. It is indeed at this primordial level that the phenomenon of “memory places” is constituted, before they become a reference for historical knowledge. These memory places function for the most part after the manner of reminders, offering in turn a support for failing memory, a struggle in the war against forgetting, even the silent plea of dead memory. These places “remain” as inscriptions, monuments, potentially as documents,41 whereas memories transmitted only along the oral path fly away as do the words themselves. It is also due to this kinship between memories and places that the sort of ars memoriae that we will discuss at the beginning of the next study was able to be constructed as a method of “loci.”
This tie between memory and place results in a difficult problem that takes shape at the crossroads of memory and history, which is also geography. This is the problem of the degree of originality of the phenomenon of dating, in parallel with localization. Dating and localization constitute in this respect solidary phenomena, testifying to the inseparable tie between the problematics of time and space. The problem is the following: up to what point can a phenomenology of dating and localization be constituted without borrowing from the objective knowledge of geometrical—let us say, Euclidian and Cartesian—space and from the objective knowledge of chronological time, itself articulated in terms of physical movement? This is the question posed by all the attempts to recover an earlier Lebenswelt—conceptually, if not historically—in the world (re)constructed by the sciences of nature. Bergson himself, so vigilant regarding the threats of contamination of the pure experience of duration by spatial categories, did not refrain from characterizing recollection-memory by the phenomenon of dating in contrast to habit-memory. Concerning particular readings, whose evocation interrupts the recitation of a lesson, he says: “It is like an event in my life; its essence is to bear a date, and consequently to be unable to occur again” (Matter and Memory, 90); and a little later, “confronted by two different memories theoretically independent,” he notes: “The first records, in the form of memory-images, all the events of our daily life as they occur in time; it neglects no detail; it leaves to each fact, to each gesture, its place and date” (92). The date, as a place in time, thus appears to contribute to the first polarization of mnemonic phenomena divided between habit and memory properly speaking. It is equally constitutive of the reflective phase, or as we have called it, the declarative phase of remembering; the effort of memory is in large part an effort of dating: When? How long ago? How long did it last? Nor did Husserl escape this question, long before the period of the Krisis, as early as The Phenomenology of the Consciousness of Internal Time. I cannot say that a sound begins, continues, stops, without saying how long it lasts. What is more, to say that B follows A is to recognize a primordial character in the succession of two distinct phenomena: “The consciousness of succession is consciousness that gives its object originally: it is ‘perception’ of this succession” (§18, 44). We are not far from Aristotle, for whom the distinction of before and after is the distinguishing factor of time in relation to movement. The consciousness of internal time as original already possesses, according to Husserl, the a priori that governs its apprehension.
Returning to the memory of places, we can attempt, following Casey, to recover the sense of spatiality on the basis of the abstract conception of geometrical space. For the latter, he employs the term “site” and reserves “place” for lived spatiality. The place, he says, is not indifferent with regard to the “thing” that occupies it or rather fills it, in the manner in which, according to Aristotle, the place constitutes what is contained within a specific volume. Some of these remarkable places are said to be memorable. The act of inhabiting, mentioned above, constitutes in this respect the strongest human tie between the date and the place. Places inhabited are memorable par excellence. Declarative memory enjoys evoking them and recounting them, so attached to them is memory. As for our movements, the successive places we have passed through serve as reminders of the episodes that have taken place there. They appear to us after the fact as hospitable or inhospitable, in a word, as habitable.
The question will, nevertheless, arise at the beginning of the second part, at the turning point from memory to history, regarding whether a historical time, a geographical space can be conceived without the help of the mixed categories that join lived time and lived space to objective time and geometrical space, which the epoché has methodically bracketed to the benefit of a “pure” phenomenology.
The question already encountered several times as to whether the Husserlian epoché is ultimately tenable arises again here. Regardless of the ultimate destiny of the memory of dates and places on the level of historical knowledge, what primordially legitimizes the disengagement of space and time from their objectified forms is the tie linking corporeal memory to the memory of places. In this regard, the body constitutes the primordial place, the here in relation to which all other places are there. The symmetry is complete in this respect between spatiality and temporality: “here” and “now” occupy the same rank, alongside “me,” “you,” “he,” and “she,” among the deictic forms that punctuate our language. Here and now, in truth, constitute absolute places and dates. But how long can we maintain this bracketing of objectified time and space? Can I avoid relating my here to the there delimited by the body of the other without having recourse to a system of neutral places? The phenomenology of the memory of places seems to be caught, from the outset, in an insurmountable dialectical movement of disinvolvement of lived space with regard to geometrical space and of reinvolvement of each by the other in every process by which what is one’s own is related to what is foreign. Could I consider myself as someone’s neighbor without a topographical sketch? And could the here and the there stand out against the horizon of a common world, if the chain of concrete neighborhoods was not set within the grid of a great cadastre in which places are more than sites? The most memorable places would not seem to be capable of exercising their memorial function if they were not also notable sites at the intersection point of landscape and geography. In short, would the places of memory be the guardians of personal and collective memory if they did not remain “in their place,” in the twofold sense of place and of site?
The difficulty referred to here becomes especially troublesome when, following Casey, we place the mnemonic phenomena tied to commemoration at the end of the path held to lead memory away from its “mentalist” core. To be sure, it is perfectly legitimate to place commemoration back within the framework of the reflexivity/worldliness polarity.42 But then the price to pay for inserting commemoration within the context of worldliness is particularly high: once the emphasis has been placed on corporeal gestures and on the spatiality of the rituals that accompany the temporal rhythms of celebration, then the question of the nature of the space and the time in which these festive figures of memory unfold cannot be avoided. Could the public space at the heart of which the celebrants are gathered together and the calendar of feasts that mark the high points of ecclesiastical liturgies and patriotic celebrations be said to fulfill their functions of assembling the community (religio equivalent to religare?) without the articulation of phenomenological space and time onto cosmological space and time? More particularly, are not the founding events and actions, ordinarily situated in a far distant time, tied to calendar time, to the extent that they sometimes determine the zero point of the official system of dating?43 An even more radical question: does not the sort of perennialization resulting from the series of ritual reenactments, continuing beyond the deaths one by one of the co-celebrants, make our commemorations the most wildly desperate act to resist forgetfulness in its most surreptitious form of erasing traces, of grinding into dust? Now this forgetfulness seems to operate at the point of intersection of time and physical movement, at the point where, Aristotle notes in Physics 4.12, time “wastes things away.” It is on this note of hesitation that I interrupt, rather than complete, this sketch of a phenomenology of memory.
Under this title “Memories and Images” we reach the critical point of the entire phenomenology of memory. It is no longer a question of a polarity capable of being embraced by a generic concept such as memory, even when it is split into the simple presence of a memory—Greek mnēmē—and recall, recollection—Greek anamnēsis. The troublesome question is the following: is a memory a sort of image, and if so, what sort? And if it should prove possible through appropriate eidetic analysis to account for the essential difference between images and memories, how could their interconnectedness, even their confusion, be explained not only on the level of language but on the level of actual experience: Do we not speak of what we remember, even of memory as an image we have of the past? The problem is not new: Western philosophy inherited it from the Greeks and from their variations on the term eikōn. To be sure, we have stated repeatedly that imagination and memory have as a common trait the presence of the absent and as a differential trait, on the one hand, the bracketing of any positing of reality and the vision of something unreal and, on the other, the positing of an earlier reality. And yet our most difficult analyses will be devoted to reestablishing the lines of transference from one problematic to the other. After having uncoupled imagination from memory, what necessity compels us to reassociate them for a reason other than that which presided over their dissociation? In a word: what is the eidetic necessity attested by the expression memory-image that continues to haunt our phenomenology of memory and that will return in full force on the epistemological level in the historiographical operation that constitutes the historian’s representation of the past?44
We will take Husserl as our first guide in the investigation of the eidetic differences between image and memory. Husserl’s contribution to this discussion is considerable, although his fragmentary analyses scattered over more than twenty-five years did not result in a finished work. Several of these analyses, however, have been collected in volume 23 of Husserliana under the title Phantasie, Bildbewusstein, Erinnerung 1898–1925,45 employing a vocabulary imposed by the state of the discussion at the end of the nineteenth century around thinkers as important as Brentano. For my part, I salute in these analyses, with their combined patience and intellectual honesty, the second major contribution of descriptive phenomenology to the problematic of memory, alongside the analyses devoted to retention and recollection in the first two sections of the 1905 “Lectures on the Phenomenology of the Consciousness of Internal Time.” It is indeed to the correlation between these two parallel series that I wish to draw the reader’s attention: each of them has to do with the “objective” side of Erinnerung which is appropriately designated in French by the substantive souvenir (memory).
These laborious texts explore the specific differences that distinguish by means of their “objective” (Gegenständlichen) correlates a variety of acts of consciousness characterized by their specific intentionality. The difficulty of the description comes not only from the interweaving of these correlates but from the linguistic burdens of prior usages, some highly traditional ones such as the use of the term Vorstellung, imperatively but unhappily translated in French (as in English) by “representation,” some others imposed by the discussions of that period. Hence the word Vorstellung, unavoidable since Kant, includes all the correlates of sensory, intuitive acts, distinct from judgment: a phenomenology of reason, which Husserl continually projected, could not do without it. But the comparison with perception and all the other intuitive sensory acts offered a more promising entry. And this is what Husserl obstinately pursued: it forced him to distinguish among a variety of “the modes of presentation” of something, perception constituting “presentation pure and simple,” Gegenwärtigung, all the other acts being classified under the heading of presentification, Vergegenwärtigung (a term also translated by “re-presentation,” at the risk of confusing re-presentation and representation, Vorstellung).
The title of Husserl’s volume covers the field of a phenomenology of intuitive presentifications. We see where the overlap can be made with the phenomenology of memory: the latter is a sort of intuitive presentification having to do with time. Husserl often places his program under the aegis of a “phenomenology of perception, of Bild, of Phantasie, of time, of the thing [Ding],” a phenomenology that has yet to be realized. The fact that perception and its mode of presentation are taken as guidelines should not prematurely give rise to a suspicion of some sort of “metaphysics of presence”—it is a matter of the presentation of something with its distinctive character of intuitivity. All the manuscripts in the volume have to do, therefore, with objective modes that share in intuitivity but differ from perception by the non-presentation of their object. This is their common feature. Their differences come later. As concerns the place of memories on this palette, it remains incompletely determined as long as its tie with the consciousness of time has not been established; but this tie can be made on the level of the analyses of retention and of reproduction that remain within the objective dimension. We must then compare, as Husserl requests, the manuscripts collected in Husserliana, vol. 10, “The Consciousness of Internal Time,” and those of volume 23. In the latter collection, what matters is the kinship with the other modalities of presentification. The stakes of the analysis at this stage concern the relation between memory and image, our word “image” occupying the same ground as Husserl’s Vergegenwärtigung. But was this not already the case with the Greek eikōn and its run-ins with phantasia? We will return to this with Bild and Phantasie. In fact, memories are involved in these two modalities, as their enumeration in Husserl’s preferred title reminds us, and to them should be added expectation (Erwartung), placed on the same side as memory but at the opposite end of the palette of temporal presentifications, as we also see in the manuscripts on time.
When Husserl speaks of Bild, he is thinking of presentifications that depict something in an indirect manner: portraits, paintings, statues, photographs, and so on. Aristotle had begun this phenomenology by noting that a picture, a painting could be read as a present image or as an image designating something unreal or absent.46 Everyday language, quite imprecise, speaks in this situation of image as well as representation; but it sometimes specifies by asking what a particular picture represents, of what it is the image. One could then translate Bild as depiction, based on the model of the verb to depict.
When Husserl speaks of Phantasie, he is thinking of fairies, angels, and devils in stories: it is indeed a matter of fiction (some texts state Fiktum). Husserl is, moreover, interested in this by reason of the ties to spontaneity, which is a feature of belief (a term he uses often in accordance with the usage of the English-language tradition). The phenomenology of memory is implied in these distinctions and these ramifications. But the examples proposed by no means eliminate the need for an essential, eidetic, analysis. And Husserl’s interminable analyses attest to the difficulty of stabilizing meanings that continue to tread one upon another.
It is the distinction between Bild and Phantasie that proved troublesome for him from the beginning (1898–1906), hence at the time of the Logical Investigations, in the context of a theory of judgment and of the new theory of meanings that pushes to the forefront the question of intuition in terms of Erfühlung, of the “fulfillment” of signifying intentions. Later, during the period of the Ideen, it is the modality of neutrality specific to Phantasie that will move to the fore, confronting the positional character of perception. Intervening as well, though obliquely, will be the question of the individuation of something, performed by the different types of presentations, as if periodically it was intuition that reasserted itself at the top of the scale of knowledge. At other times, it is the extreme distancing of Phantasie in relation to presentation in the flesh that intrigues him. Phantasia then tends to occupy the entire place held by the English word “idea” as it is opposed to “impression” in the British empiricists. It is no longer simply a matter of devilish intrigues but also of poetic or other fictions. It is non-presenting intuition that delimits the field. Should we venture to speak tranquilly of fantasy, of the fantastic in the manner of the Greeks? (The graphism “phantasy” or “fantasy” then remains open.) What matters to the phenomenology of memory is that the temporal note of retention can be linked up with fantasy considered provisionally as a genus common to all non-presentations. However, the vocabulary of Vorstellung is retained when the emphasis falls on the intuition common to presentation and to presentification in the field of a phenomenological logic of meanings. Is it then on Phantasie alone that the temporal marks of retention and reproduction are to be grafted? Yes, if the emphasis falls on non-presentation. No, if it falls, in the case of secondary remembering, on reproduction: then the kinship with Bild is imposed, which, beyond the examples mentioned above, covers the entire field of the “depicted” (das Abgebildete), that is to say, of an indirect presentification based on a thing itself presented. And if the emphasis falls on “the belief of being attached to the memory” (Seinsglaube an das Erinnerte), then the opposition between memory and fantasy is complete: the latter lacks the present “as it were” of the reproduced past. On the other hand, the kinship with the “depicted” seems more direct, as when one recognizes a loved one in a photograph. The “remembered” then draws upon the “depicted.” It is with this play of attractions and repulsions that Husserl continues to struggle.47 The sole fixed point remains the theme of intuitive presentifications, taking into account their own entanglement with the conceptual modalities of representation in general, a theme that covers presentations and non-presentations, hence the totality of objectifying “apprehensions,” leaving out only practical and affective lived experiences, which, in truth, were presumed to be constructed on the basis of these apprehensions.
The field thus continues at times to widen to include all Auffassungen (“apprehensions”), at times to narrow to the innumerable ramifications of presentifications or representations. The interplay between the remembered, the fictive (Fiktum) and the depicted (Abgebildete) is then invoked against the backdrop of the global opposition to perception, whose object presents itself directly (Selbstgegenwärtige); the depicted advancing over the pretended by its indirect character, a physical image (Bild) offering support. The split then passes between the image (Bild) and the thing (Sache in the sense of res, pragmata), the thing in question, not the thing (Ding) in space.
If a memory is an image in this sense, it contains a positional dimension that, from this point of view, brings it closer to perception. In another vocabulary, which I am adopting, one speaks of the having-been of the remembered past, the ultimate referent of the memory in action. What will then pass to the forefront, from the phenomenological point of view, will be the split between the unreal and the real (whether it be present, past, or future). While imagination can play with fictional entities, when it does not depict but cuts itself off from the real, memories posit past things; whereas the depicted still has one foot in presentation as indirect presentation, fiction and the pretend are situated radically outside of presentation. However, considering the diversity of viewpoints under which phenomena are described and the variable scope recognized concerning these phenomenological types, “consciousness of Bild” and “consciousness of Phantasie” can, in turn, be distinguished from one another on the same plane, then set in opposition to one another, or made to include one another in one sense or in the other, depending on the place that is given to them in the field of intuitive presentifications: the entire place or part of it. (It happens that Husserl reserves the substantive Phantasma for these supports for the operation of depicting, pulling Phantasie itself in this way into the field of depicting the Bild.)48
It is this inclusive problematic of presentification that will be upset in the third section of The Phenomenology of the Consciousness of Internal Time. The opposition between presentation and presentification continues, nonetheless, to function within the objective field of the correlates of intentional consciousness, as does the distinction between primary memory and secondary memory, considered as temporal varieties of presentification, of “making present” that which does not give itself as present in the sense of presenting. The same analyses made on the basis of memories and no longer concerning Bild or Phantasie add to the complexity. As past, the thing remembered would be a pure Phantasie,49 but, as given once more, it imposes memory as a modification sui generis applied to perception;50 under this second aspect Phantasie would place a memory in “suspension” (aufgehoben),51 which would make the memory simpler than the fiction. We would then have the sequence: perception, memory, fiction. A threshold of inactuality is crossed between memory and fiction. The phenomenology of memory therefore has to free itself from the tutelage of fantasy, of the fantastic, marked with the seal of inactuality, of neutrality. Yet to evoke neutrality, as was done in Ideen I, §111, in order to situate the fantastic in relation to the remembered, is to invoke belief: to the certainty common to the series: perception, memory, expectation is opposed a mode of uncertainty such as admission (Aufnahme), presentiment (Ahnung). These modalities belong to the same cycle as do all “positings” (Stellungnahmungen), the genus common to all the modalities of the inactual, the neutral.
The dividing line thus runs all along the break between presentation and presentification. Memories are a specific modification of presentation, at least as primary memory or retention, as confirmed by the first section of the 1905 lectures. Here Husserliana, vol. 23 and Husserliana, vol. 10 coincide, their primary emphasis bearing on the operative mode (or performance) (Vollzug), which distinguishes reproduction from production, inactuality from actuality, non-positing from positing. Any possibility of confusing a memory with an image in the sense of the term Bild is henceforth eliminated. Everything is played out on the scene of the “objective” correlate of the experiences interrogated.
Ideen I, despite the idealist turn taken by the philosophy of consciousness, will not speak a different language concerning the “manner of fulfillment” of the intuitive modalities included within the scope of presentification.52 The criterion of positionality will continue to be strengthened in the texts coming after Ideen I: memories belong to the “world of experience” in contrast to the “worlds of fantasy,” of irreality. The former is a common world (without, as yet, any mention of the manner of intersubjective mediation), the latter are totally “free,” their horizon completely “undetermined.” In principle, then, they cannot be confused or mistaken one for the other, whatever may be said regarding the complex relations between Fiktum and possibility, even their irreducibility to one another. A phenomenology attentive to eidetic differences never finishes making distinctions.
If one had to define the difference in approach between the applications in Husserliana, vol. 10 (which themselves repeat those of the first section of the 1905 “Lectures on the Phenomenology of the Consciousness of Internal Time”) and the applications concerning the sequence Phantasie, Bild, Erinnerung, one could say that with regard to the latter the emphasis is placed on the differences between members of the family of presentifications, hence on the modifications affecting the presentations of the “objective” correlate, while in the 1905 lectures, the emphasis falls on the temporal modalities specific to the sort of presentification characterizing memories. In this respect, it is noteworthy that in the analyses in Husserliana, vol. 23 the key notion of presentation (Gegenwärtigung) is still distinguished from the temporal present, just as the theme of the now (Jetzt) is still absent from the objective analysis of memories, without producing any negative effects. Must we not conclude that the present—the now, a notion by which the series of indicators of temporality are to be governed—is not to be separated from the idea of presentation, to which the various types of presentification are themselves referred? And if this hypothesis has merit, is it not then the kinship between memories and images within the great family of presentifications that authorizes, retrospectively, the break I made when I stopped the movement that carries the entire work of the 1905 lectures toward the self-constitution of the flow of consciousness at the objective moment? The transition will turn on the return to the self, from intentionality ad extra—transversal, as it is called—still at work in the phenomenology of memory, to intentionality ad intra, or longitudinal, which predominates in the self-constitution of the flow. We will retie this broken thread in the third chapter of the phenomenology of memory.
At the end of this voyage in the company of Husserl through the labyrinth of entanglements that make this peregrination a difficult one, I must confess that only half of the route has been covered when we account for the confusion that hampers the comparison between image and memory. How are we to explain that memories return in the form of images and that the imagination mobilized in this way comes to take on forms that escape the function of the unreal? It is this double imbroglio that we must now untangle.
I am adopting here as my working hypothesis the Bergsonian conception of the passage from “pure memory” to memory-image. I am speaking of a working hypothesis, not to separate myself from his fine analysis but from the outset to indicate my concern with distinguishing in the text of Matter and Memory, as far as this is possible, the psychological description from the metaphysical (in the strong and noble sense of the word) thesis concerning the role assigned to the body and to the brain and, consequently, asserting the immateriality of memory. This bracketing of the metaphysical thesis amounts to dissociating, in the heritage received from the Greeks, the notion of eikōn from that of tupos, of the imprint, which was associated with it from the start. From the phenomenological point of view, the two notions belong to two distinct orders: the eikōn contains within itself the other of the original affection, while the tupos involves the external causality of an impetus (kinēsis), which is itself at the origin of pressing the seal into the wax. The entire modern problematic of “mnemonic traces” is, in fact, heir to this ancient alliance between eikōn and tupos. The metaphysics of Matter and Memory proposes, precisely, to link systematically the relation between the action, the center of which is the brain, and the pure representation which is self-sufficient as a result of the persistence, in principle, of the memory of the initial impressions. It is this presumed relation that I am bracketing in the analysis that follows.53
The distinction Bergson makes between “pure memory” and memory-image radicalizes the thesis of the two memories with which we began the preceding phenomenological sketch. And it is, therefore, this thesis that is made even more radical in its turn by the metaphysical thesis upon which Matter and Memory is constructed. It is within this intermediary location, with regard to the strategy of the work as a whole, that we will carry through our description of the passage from “pure memory” to memory-image.
Let us start the analysis by accepting that there does exist something like a “pure memory” that has not yet been put into images. We will say a little further on in what way it is possible to speak of this and how important it is to be able to speak convincingly of it. Let us start at the furthest point reached by the theory of the two memories: “To call up the past in the form of an image, we must be able to withdraw ourselves from the action of the moment, we must have the power to value the useless, we must have the will to dream. Man alone is capable of such an effort. But even in him the past to which he returns is fugitive, ever on the point of escaping him, as though his backward turning memory were thwarted by the other, more natural, memory, of which the forward movement bears him on to action and to life” (Bergson, Matter and Memory, 94). At this stage of the analysis, we have available to us in speaking of “pure memory” only the example of a lesson learned by heart. And it is by a sort of passage to the limit that we write, following Bergson: “Spontaneous recollection is perfect from the outset; time can add nothing to its image without disfiguring it; it retains in memory its place and date” (95). The distinction between a “memory which recalls” and a “memory which repeats” was the fruit of a method of division that consisted in distinguishing “two extreme forms of memory in their pure state,” then in reconstructing the memory-image as an intermediary form, as a “mixed phenomenon which results from their coalescence” (103). And it was in the act of recognition that this fusion occurred, signaled by the feeling of déjà vu. It is also then in the work of recollection that this operation of putting the “pure memory” into images can be grasped in its origin. We can speak of this operation only as a movement from the virtual to the actual, or again as the condensation of a cloud or as the materialization of an ethereal phenomenon. Other metaphors suggest themselves: movement from the depths to the surface, from shadows to the light, from tension to relaxation, from the heights to the lower levels of psychical life. Such is the “movement of memory at work” (171). It carries memory back so to speak into a region of presence similar to that of perception. But—and here we reach the other side of the difficulty—it is not just any sort of imagination that is mobilized. In contrast to the function of derealization, culminating in a fiction exiled to the margins of reality considered in its totality, what is celebrated here is instead the visualizing function of imagination, its manner of giving something to be seen. On this point, what unavoidably comes to mind is the final component of the muthos that, according to Aristotle’s Poetics, structures the configuration of tragedy and epic, namely, the opsis, held to consist in “placing before the eyes,” showing, making visible.54 This is also the case when “pure memory” is put into images: “Essentially virtual, it cannot be known as something past unless we follow and adopt the movement by which it expands into a present image, thus emerging from obscurity into the light of day” (173). The strength of Bergson’s analysis is to keep the two extremities of the spectrum separate and yet connected. At one end: “To imagine is not to remember. No doubt a recollection, as it becomes actual, tends to live in an image; but the converse is not true, and the image, pure and simple, will not be referred to the past unless, indeed, it was in the past that I sought it, thus following the continuous progress which brought it from darkness into light” (173–74, trans. modified).
If we follow this thought to the other extreme, descending from “pure memory” to memory-image—and, as we shall see, far beyond that—we witness a complete reversal of the imaging function, whose shadow also extends from the far pole of fiction to the opposite pole of hallucination.
It was the fiction-pole of the imagination that I considered in Time and Narrative when I opposed fictional narrative to historical narrative. It is in relation to the other pole, the hallucination-pole, that we now have to situate ourselves. Just as Bergson dramatized the problem of memory by his method of division and the shift to opposing poles, it is important to dramatize the thematic of the imagination in the same way by organizing it in relation to the two poles of fiction and hallucination. By moving to the pole of hallucination, we uncover the pitfall of the imaginary for memory. It is, in fact, just this sort of memory that is a common target of the rationalist critiques of memory.
In order to account for this trap, I thought it might be appropriate to summon, alongside Bergson, another witness, Jean-Paul Sartre in The Psychology of Imagination.55 This astonishing book sets off along the path of just such a reversal of the problematic of memory, even though this is not its purpose. I called this book astonishing. It begins, in fact, with a plea for a phenomenology of the unreal, approaching from the other side the effort of uncoupling imagination and memory, which we attempted earlier. As is firmly asserted in the conclusion, despite the drift we will discuss: “the hypothesis of the imaginative consciousness is radically different from the hypothesis of a consciousness of the real. This means that the type of existence of the object of the image, as long as it is imagined, differs in nature from the hypothesis of existence of the object of the real. . . . This essential nothingness of the imagined object is enough to distinguish it from the object of perception” (261). Memory is on the side of perception, as concerns its thesis of reality: “there is . . . an essential difference between the theme of a recollection and that of an image. If I recall an incident of my past life I do not imagine it, I recall it. That is, I do not posit it as given-in-its-absence but as given-now-in-the-past in the past” (263). This is exactly the interpretation proposed at the beginning of this study. But now here is the reversal. It takes place on the terrain of the imaginary. It results from what can be called the hallucinatory seduction of the imaginary. The fourth part of The Psychology of Imagination is devoted to this seduction under the title “The Imaginary Life”: “The act of imagination . . . magic alone. It is an incantation destined to produce the object of one’s thought, the thing one desires, in a manner that one can take possession of it” (177). The incantation is equivalent to the voiding of absence and distance. “This is a way of playing at satisfying my desire” (179). The imagined object’s “not-being-there” is covered over by the quasi presence induced by the magical operation. Its unreality is warded off by this sort of “dance before the unreal” (205). In truth, this voiding was nascent in “placing before the eyes” considered as “putting into images,” the putting-on-stage constitutive of the memory-image. In this text, Sartre did not foresee the rebound effect on the theory of memory. But he paves the way for this understanding in his description of what is soon to become a “pathology of the imagination” (213ff.). It is centered on the hallucination and its distinctive feature, obsession, namely, “that sort of vertigo inspired in particular by flight in the face of that which is forbidden.” Every effort “not to think about it anymore” is spontaneously transformed into “obsessive thinking.” Confronting the phenomenon of fascination with the forbidden object, how can we help but leap to the plane of collective memory and evoke the sort of hauntedness, described by historians of the present day, which stigmatizes this “past that does not pass”? Hauntedness is to collective memory what hallucination is to private memory, a pathological modality of the incrustation of the past at the heart of the present, which acts as a counterweight to the innocent habit-memory, which also inhabits the present, but in order to “act it” as Bergson says, not to haunt it or torment it.
From Sartre’s description of the reversal of the imagination’s function of derealization into a function of hallucination, a curious parallel results between the phenomenology of memory and the phenomenology of imagination. It seems as though the form that Bergson calls intermediary or mixed memory—namely, memory-image, half-way between “pure memory” and memory reinscribed in perception, at the stage where recognition blossoms in the feeling of déjà-vu—corresponded to an intermediary form of imagination, half-way between fiction and hallucination, namely, the “image” component of the memory-image. So it is also as a mixed form that we must speak of the function of the imagination consisting in “placing before the eyes,” a function that can be termed ostensive: this is an imagination that shows, gives to be seen, makes visible.
A phenomenology of memory cannot fail to recognize what we have just called the pitfall of the imaginary, inasmuch as this putting-into-images, bordering on the hallucinatory function of imagination, constitutes a sort of weakness, a discredit, a loss of reliability for memory. We will return to this when we consider a certain way of writing history, after the manner of Michelet, we might say, in which the “resurrection” of the past also tends to take on quasi-hallucinatory forms. In this way, writing history shares the adventures of memories put-into-images under the aegis of the ostensive function of imagination.
I do not want to conclude on this note of perplexity, but instead with the provisional response than can be given to the question of trust that the theory of memory passes on to the theory of history. This is the question of the reliability of memory and, in this sense, of its truth. This question stood in the background of our entire investigation concerning the differential feature that separates memory from imagination. At the end of our investigation, and in spite of the traps that imagination lays for memory, it can be affirmed that a specific search for truth is implied in the intending of the past “thing,” of what was formerly seen, heard, experienced, learned. This search for truth determines memory as a cognitive issue. More precisely, in the moment of recognition, in which the effort of recollection is completed, this search for truth declares itself. We then feel and indeed know that something has happened, something has taken place, which implicated us as agents, as patients, as witnesses. Let us call this search for truth, faithfulness. From now on, we will speak of the faithfulness of memories, of memories being true to . . . , in order to express this search, this demand, this claim, which constitutes the veridical-epistemic dimension of the orthos logos of memory. The study that follows will have the task of showing how the epistemic, veridical dimension of memory is united with the practical dimension tied to the idea of the exercise of memory.