Science Fiction Studies

#4= Volume1, No. 4 = Fall 1974

Stanislaw Lem

Todorov's Fantastic Theory of Literature

Translated from the Polish by Robert Abernathy

Since structuralism in literary studies is largely of French origin, let this attempt to ruin its reputation have as its motto the words of a Frenchman, Pierre Bertaux:

At one time it was hoped that the beginnings of a formalization of the humanities analogous [to that of the "diagonal" or "formalistic" sciences] could be expected from structuralism. Unfortunately it appears today that precisely the loudest advocates of structuralism have let it degenerate into a mythology-- and not even a useful one. This chatter that is now called structuralism has apparently dealt a mortal blow to that rudimentary scientific beginning.1

I fully agree with this verdict. However, inasmuch as it is difficult to expose in a single article the barrenness of a whole school of thought-one moreover which has spawned divergent tendencies, since here every author has his own "vision" of the subject-I will limit myself to dissecting Tzvetan Todorov's book, The Fantastic.2

THE HISTORY OF THE DEGENERATION of a conceptual apparatus that originated in mathematical linguistics, after it was mechanically transplanted into the domain of metaliterature, has yet to be written. It will show how defenseless logical concepts become when they are torn out of contexts in which they were operationally justified, how easy it is, by parasitizing on science properly speaking, to bemuse humanists with pretentious claptrap, disguising one's actual powerlessness in a foreign field beneath a putatively unassailable logical precision. This will be a rather grim, but instructive, history of how unambiguous concepts turn into foggy ones, formal necessity into arbitrariness, syllogisms into paralogisms. It will, in short, deal with a retrograde trend in French critical thought, which, aiming at nothing less than logical infallibility in theory-building, transformed itself into an incorrigible dogmatism.

Structuralism was to be a remedy for the immaturity of the humanities as manifested in their lack of sovereign criteria for deciding the truth or falsehood of theoretical generalizations. The formal structures of linguistics are mathematical in origin, and are, indeed, numerous and diverse, corresponding to branches of both pure and classical mathematics ranging from probability and set theories to the theory of algorithms. The inadequacy of all these leads linguists to employ new models, e.g. from the theory of games, since this furnishes models of conflicts, and language is, at its higher, semantic levels, entangled in irreducible contradictions. These important tidings have, however, not yet reached those literary scholars who have taken over a small fraction of the arsenal of linguistics and endeavor to model literary works using conflict-free deductive structures of an uncommonly primitive type-- as we shall demonstrate on the example of the Todorov book.

THIS AUTHOR BEGINS BY DISPOSING of some objections which arise in connection with constructing a theory of literary genres. Deriding the investigator who would, before proceeding to description of a genre, engage in endless reading of actual works, he asserts—appealing to the authority of Karl Popper—that for the maker of generalizations it suffices to be acquainted with a representative sample from the set of objects to be studied. Popper, wrongly invoked, is in no wise to blame, since representativeness of a sample in the natural sciences and in the arts are two quite different matters. Every normal tiger is representative for that species of cats, but there is no such thing as a "normal story." The "normalization" of tigers is effected by natural selection, so the taxonomist need not (indeed should not) evaluate these cats critically. But a student of literature who is in like fashion axiologically neutral is a blind man confronting a rainbow, for, whereas there do not exist any good organisms as distinguished from bad ones, there do exist good and worthless books. And in the event, Todorov's "sample," as displayed in his bibliography, is astonishing. Among its twenty-seven titles we find no Borges, no Verne, no Wells, nothing from modern fantasy, and all of SF is represented by two short stories; we get, instead, E.T.A. Hoffmann, Potocki, Balzac, Poe, Gogol, Kafka—and that is about all. In addition, there are two crime-story authors.

Todorov declares, further, that he will pass over problems of esthetics altogether in silence, since these are beyond the present reach of his method.

Thirdly, he debates the relationship of the Species and its Specimen. In nature, he says, the occurrence of a mutation does not modify the species: knowing the species tiger, we can deduce from it the properties of each individual tiger. The feedback effect of mutations upon the species is so slow that it can be ignored. In art it is different: here every new work alters the species as it existed heretofore, and is a work of art just insofar as it departs from a specific model. Works which do not satisfy this condition belong to popular or mass literature, such as detective stories, slushy love stories, SF, etc. Agreeing thus far with Todorov, I see what is in store for his method as a result of this state of affairs: the more inferior and paradigmatically petrified the texts which it undertakes to anatomize, the more readily it will reveal structures. Todorov not surprisingly, omits to draw this conclusion.

Further, he discusses the question of whether one should investigate genres which have arisen historically or those which are theoretically possible. The latter strike me as coming to the same thing as a history of mankind a priori, but since it is easier to formulate a foolish idea concisely than it is to refute it concisely, I will let this pass. I will however remark here that there is a difference between taxonomy in nature and in culture which structuralism overlooks. The naturalist's acts of classification, say of insects or of vertebrates, evoke no reaction on the part of that which is classified. A futurologist might say that Linnaean taxonomy is not subject to the Oedipus effect (Oedipus got into trouble by reacting to a diagnosis of his fate). On the other hand, the literary scholar's acts of classification are feedback-linked to that which is classified, i.e. the Oedipus effect manifests itself in literature. Not straightforwardly, to be sure. It is not the case that writers, upon reading a new theory of genres, run straight to their studios to refute it by means of their next books. The linkage is more roundabout. Sclerosis of paradigms, as a stiffening of intergeneric barriers, arouses authors to a reaction which expresses itself, among other ways, in the hybridization of genres and the attack on traditional norms. Theoreticians’ labors are a catalyst which accelerates this process, since their generalizations make it easier for writers to grasp the entire space of creative activity, with its inherent limitations. Thus the student of genres who establishes their boundaries causes writers to rebel against them—he produces a feedback loop by the very act of classification. To describe limitations on creativity thus amounts to drawing up a self-defeating prognosis. What could be more tempting than to write what theory prohibits?

The constriction of the imagination which is inherent in a dogmatic mentality, such as is represented by the structuralist, manifests itself in the belief that what he has found to be barriers to creativity can never be transgressed by anyone. Perhaps there exist untransgressible structures of creativity, but structuralism has not come within reach of any such. Rather, what it proclaims to us as bounds of creativity is really quite an antique piece of furniture, to wit the bed of Procrustes, as we shall show.

COMING TO MATTERS OF SUBSTANCE, Todorov first of all demolishes past attempts at defining the fantastic. After crossing off the efforts of Northrop Frye, he lights into Roger Caillois, who had the bad luck to write that a "touchstone of the fantastic" is "the impression of irreducible strangeness" (p. 35). According to Caillois, jeers Todorov, a work's genre depends on the sang-froid of its reader: if he is frightened, then we have to do with the (uncanny) fantastic, but if he keeps his presence of mind, then the work must needs be reclassified from the standpoint of the theory of genres. We will speak in the proper place of how the scoffer has here left his own method exposed to attack.

Todorov distinguishes three aspects of the literary work: the verbal, the syntactic and the semantic, making no secret of the fact that these were formerly known as style, composition and theme. But their invariants have traditionally and mistakenly been sought "on the surface" of texts; Todorov declares that he will look for structures on a deep level, as abstract relations. Northrop Frye, suggests Todorov, might say that the forest and the sea form a manifestation of an elementary structure. Not so—these two phenomena manifest an abstract structure of the type of the relation between statics and dynamics. Here we first come upon the fruits of spurious methodological sophistication, that congenital trait of structuralism, for it is plain to see what our author is seeking: oppositions which come to light on a level of high abstraction. Now, this one is wide of the mark, because statics is not opposed to dynamics but is a special case of it, namely a limiting case. This is a small matter, but a weighty problem lies behind it, since it is in the same way that Todorov constructs his integral structure for fantastic literature. This, by the structuralist's decree, consists of a one-dimensional axis, along which are situated sub-genres that are mutually exclusive in a logical sense. This is portrayed by Todorov's diagram: "uncanny : fantastic-uncanny : fantastic-marvelous: marvelous" (p. 44).

What is the "fantastic"? It is, Todorov explains, the hesitation of a being who knows only natural laws in the face of the supernatural. In other words, the fantastic character of a text resides in a transient and volatile state during the reading of it, one of indecision as to whether the narrative belongs to a natural or a supernatural order of things.

The "pure" uncanny amazes, shocks, terrifies, but does not give rise to indecision (of the kind which we would call ontological). This is the place of the horror story which presents occurrences that are frightful, extraordinary, but nevertheless rationally possible. This genre extends off the diagram to the left, merging into "ordinary" literature-as a transitional link our theoretician mentions Dostoevski.

The fantastic-uncanny already gives occasion to the vacillations that evoke the sense of the fantastic. This is a tale the events in which are, as its reader at first supposes, brought about by the intervention of the Supernatural. Its epilog, however, furnishes a surprising rational explanation. (Here belongs, for example, the Manuscrit trouvé à Saragosse.)

The "fantastic-marvelous" work is just the other way round—it supplies in the end explanations of an extramundane, irrational order, as in Villiers de l'Isle-Adam's Véra, inasmuch as the conclusion of this story forces one to acknowledge that the dead woman really rose from the grave.

And finally the "pure" marvelous, which again does not give rise to any vacillations between mutually exclusive types of ontic systems, has all of four subdivisions: (a) the "hyperbolic marvelous," stemming from narrative extravagance, as in the voyages of Sinbad, where he speaks of serpents capable of swallowing elephants; (b) the "exotic marvelous": here too Sinbad serves Todorov's purpose, when he says that the Roc had legs like oak trees-this is not a zoological absurdity, since to long-ago readers such an avian form may have seemed "possible"; (c) the "instrumental marvelous"—the instruments are fabulous objects such as the lamp or the ring of Aladdin; (d) and finally the fourth type of the marvelous is constituted by the "scientific," i.e. science fiction: "these narratives, starting from irrational premises, link the ‘facts’ they contain in a perfectly logical manner" (pp. 56-57). Or. "The initial data are supernatural: robots, extraterrestrial beings, the whole interplanetary context" (p. 172). And finally: "Here the supernatural is explained in a rational manner, but according to laws which contemporary science does not acknowledge" (p. 56).

The scientific bibliography of the theory of "robots" forms a thick volume; there exists a world-renowned organization of astrophysicists (CETI) concerned with searching for signals emitted by Todorov’s "supernatural beings," i.e. by extraterrestrial creatures; for our theoretician even the "interplanetary background" possesses supernatural properties. Let us however regard all these qualifications as slips of the pen. We may as well do so, since Todorov’s theory would be fine if it contained only such defects.

As we know, Todorov calls the fantastic a transitional boundary state on an axis whose opposite extremes signify the rational system of Nature and the irrational order of marvels. For a work to manifest its fantastic character, it must be read literally, from the standpoint of naive realism, thus neither poetically nor allegorically. These two categories, according to Todorov, exclude one another with logical necessity, hence fantastic poetry or fantastic allegory is always impossible. This second categorial axis is perpendicular to the first. Let us clarify these relationships on a "microexample" of our own, given by a single simple sentence. The sentence "A black cloud swallowed the sun" can be taken, first of all, as a poetic metaphor (a thoroughly trite one, but that is beside the point). The cloud, we know, was only figuratively compared to a being capable of devouring the sun, since in fact it merely hid it from view.

Furthermore, it is possible, by dint of contextual suggestions, to substitute for the cloud, say, falsehood, and for the sun, truth. The sentence becomes an allegory: it says that falsehood may obscure truth. Again, this is a platitude, but the relations which hold are clearly apparent, and that is what we are after.

Now if instead we take the sentence literally, some uncertainties emerge which make it possible for indecision and, by the same token, the fantastic to result. The cloud, we know, "actually swallowed the sun"—but in what order of events, the natural or the marvelous? if it gulped it down as a fairy-tale dragon might, then we find ourselves in a fairy tale, in the "pure marvelous." But if it engulfed the sun as did a certain cosmic cloud in the novel The Black Cloud by the astrophysicist Fred Hoyle, we shift to SF. In this novel the cloud is made of cosmic dust, it is a "cybernetic organism" and it engulfed the sun because it feeds on stellar radiation. The explanation acquires rationality as a hypothetical extrapolation from such disciplines as the theory of self-organizing systems, the theory of evolution, etc.

To be sure, the results of our classification do not coincide with Todorov’s, since for him SF is irrationalism embodied in pseudoscience. There is no point to arguing about Hoyle’s Black Cloud. It is enough to note that SF is nourished by scientific revelations—e.g. in the aftermath of the heart transplants there appeared swarms of fictional works which described criminal gangs snatching hearts from the breasts of young people on behalf of rich oldsters. Even if this is. improbable, it assuredly does not belong to any supernatural order of things. But after all, arbitration might reconcile the conflicting viewpoints by effecting, say, within the scope of Todorov’s axis, a translation of some titles, at least, toward the pole of the Rational.

Things get worse when it comes to subgenres of the fantastic for which there is no place at all on Todorov’s axis. To what genre should Borges’ "Tres versiones de Judas" be assigned? In this work Borges invented the fictional heresy of a Scandinavian theologian, according to which Judas, not Jesus, was the true Redeemer. This is not a "marvelous" tale—no more than any genuine heresy such as the Manichean or the Pelagian. It is not an apocryphon, for an apocryphon pretends to be an authentic original, while Borges’ text does not try to conceal its literary nature. It is not an allegory, nor is it poetry, but, since nobody ever proclaimed such an apostasy, the matter cannot be placed in the order of real events. Quite obviously we have to do here with an imaginary heresy, that is with fantastic theology.

Let us generalize this interesting case. Let us recognize unprovable propositions, such as metaphysical, religious or ontological assertions, as forming an "actual religious credo," a confession of faith, the affirmation of a world view, if they have entered in just this guise into the repository of the historic civilizations. From an immanent standpoint it cannot be discerned from any such proposition whether it was uttered with the conviction that things are really as it claims, or whether it was enunciated non-seriously (in "ludic" fashion, thus non-assertively). If no philosopher named Arthur Schopenhauer had ever existed and if Borges had invented in a story a doctrine called "Die Welt als Wille und Vorstellung," we would, accept this as a bit of fiction, not of the history of philosophy. But of what kind of fiction, indeed? Of fantastic philosophy, because it was published non-assertively. Here is a literature of imaginary ideas, of fictional basic values, of other civilizations-in a word, the fantasy of the "abstract."

On Todorov’s axis there is likewise no place for fantastic history, which did not happen but might have. This is a matter of so-called PF (political fiction), telling of what might have been if Japan rather than the USA had fabricated the atomic bomb, if the Germans had won World War II, and the like. These are not uncanny tales—at any rate no more so than what has actually happened in the present century—and they are not marvelous, since it would hardly have taken a miracle to make Japanese physicists go to work building reactors, and also there is no question of the reader’s being unsure about whether the narrated events are rational or irrational—and yet in just this way objective worlds are constructed, the nonexistence of which in past, present or future is an irrefragable certainty. So what sort of books are these? Beyond a doubt, ones which fabricate a fantastic universal history.

Thus our Procrustes has not made place on his meager axis even for actually existing varieties of the fantastic-let alone "theoretically possible" kinds, for which there is a fortiori no room in his bed of torture.

LET US NOW TAKE A CLOSER LOOK at Todorov’s axis. It is of logical ancestry. The structuralist is indebted to the linguists, and they in turn adopted this simplest structure of exclusion from set theory, in that here the principle of the excluded middle holds: an element either belongs to a set or it does not, and 45% membership in a set is impossible. Todorov ascribes to this axis a fundamental, because definitional, significance on the highest level of abstraction. However, the essential thing is not the axis but the reader’s act of decision. Reading a literary work indeed calls for decisions-in fact not just one, but an ordered set of them, as the resultant of which the genre classification of the text comes about. The reader’s decisions do not oscillate in only one dimension. Assuming as a working hypothesis that these are always decisions with respect to simple (binary) alternatives, thus dichotomous, one can enumerate such additional axes as:

(a) Earnestness: irony. Irony is calling a statement in question, either its linguistic level (this has been done stylistically by Gombrowicz) or its objective level. As a rule irony is in some measure reflexive. But lest the "deflation" of the utterance should become self-destruction on its part, this tactic stabilizes the reader’s hesitancy, or renders futile the attempt at a definitive diagnosis with respect to the designated opposition. It achieves its optimum durability when the separation of an "ironic component" from a "serious component" in the text is not feasible. "Tres versiones de Judas" is of just this kind.

(b) Autonomous (reflexive) text : relative text (referred to something outside itself). Todorov’s "allegory" is a bag into which countless heterogeneous matters are stuffed. Culturally local (ethnocentric) allegory is something different from universal allegory. What is allegorical in the author’s cultural sphere may be "mere entertainment" or "pure fantasy" for ethnically alien readers, in line with the saying: "Wer den Dichter will verstehen, muss in Dichters Lande gehen." The symbolism peculiar to Japanese prose may be unrecognizable by us, for precisely this reason. And again, symbolic character of a text does not necessarily make it allegorical. Whatever is a normative symbol (pertaining to taboo, say) of a given culture is by that very fact neither arbitrary, not fantastic, nor "imaginary" for that culture’s members. Whether a given text is autonomous or relative is determined by the community of culture between the author and his readers.

(c) Text as cryptogram: text as literal message. This is a variant of the foregoing opposition. The difference between the two is that in the type (b) opposition it is a matter of relations among objects (events), but in the type (c) opposition one of linguistic relations among utterances. Allegory is a sort of generalization signaled by events-objects (a man, as by Kafka, turns into an insect). The content of a cryptogram, on the other hand, can be anything, e.g. another cryptogram. From the fact that cryptograms exist it does not follow that everything is a cryptogram. From the fact that in certain cultures a part is played by themes concealed under relationships (social, familial) it does not follow that in every culture its relational character (its structure) must be a camouflage for meanings concealed in this fashion. This is why one feels a cognitive disappointment in reading Lévi-Strauss, because one cannot discover any reason, psychological, social, or logical, responsible for some meanings’ functioning in the community in overt relationships (i.e. ones publicly called by their names), whereas others are "hidden" in the network of occurring relations and have to be reconstructed by abstraction. Here for ethnological structuralism there lies in wait the same bottomless pitfall that menaces psychoanalysis, since as in psychoanalysis it is possible to impute to the analysand’s every word the status of a "mask" concealing another, deeper content, so in structuralism it is always possible to hold that what occurs as relations in a culture is inconclusive and unimportant, because it represents a "camouflage" for other concepts, those which will only be brought to light by the abstract model. Neither of these hypotheses can be verified, so they are non-empirical with respect both to assumptions and to methods.

One could go on enumerating such oppositions. Superimposing their axes, so that they form a multidimensional "compass card," i.e. a coordinate system with multiple axes, we obtain a formal model of the situation of the reader who has to make repeated decisions about a complexly structured text. Not all texts activate the decision process along all the possible axes, but a theory of genres must take into account at least that class of decisions which cumulatively determines the genre classification of what is read.

It should be emphasized that particular decisions, until they are made, are dependent variables. Once we have concluded, for example, that a text really is ironic, we have thereby altered the probabilities of specific decisions on other axes.

The perfidy of modern creative writing lies just in making life—that is, semantic decisions—difficult for the reader. Such writing was emphatically initiated by Kafka. Todorov, unable to cope with Kafka’s texts by means of his axis, has made a virtue of methodological paralysis, taking his own perplexity out into the deep waters of hermeneutics. According to him, Kafka conferred "complete autonomy" on his text, he cut it off from the world in all directions. The text seems to be allegorical but is not, since there is no way of ascertaining to what court it addresses its appeal. Hence it is neither allegorical nor poetic nor realistic, and if it can be called fantastic, then only in the sense that "dream logic" has engulfed the narrative together with the reader. ("Son monde tout entier obéit à une logique onirique sinon cauchemardesque, qui n’a plus rien à voir avec le réel." [p. 1811) Ita dixit Todorov, without noticing that he has hereby abandoned all his structuralizing.

Todorov’s conception of Kafka’s works as totally lacking an address (as reflexive) in the real world ("n’a plus rien à voir avec le réel") has become popular also outside structuralist circles, I think, as a result of intellectual laziness. These works, boundlessly veiled in meanings, seem to signify so much at once that no one knows what they mean concretely—well, then, let it be that they simply mean nothing, whether referentially, allusively, or evocatively.

If there existed an experimental science of literature concerned with studying readers’ reactions to deliberately prepared texts, it would prove in short order that a text wholly severed from the world with regard to its meanings can be of no interest to anyone. References of expressions to extralinguistic states of affairs form a continuous spectrum, ranging from ostensive denotation to an aura of allusions hard to define, just as recall of things seen to our visual memory ranges from sharp perception in broad daylight to the vagueness of a nocturnal phantom in the dark. Consequently, a boundary between "undisguised reference" and "hermetic autonomy" of a text can be drawn only arbitrarily, because the distinction is extremely fuzzy.

A representative of impressionistic criticism might say that Kafka’s writing "shimmers with mirages of infinite meanings," but an advocate of scientific criticism must uncover the tactics which bring this state of things about, not hand the texts a charter certifying their independence of the visible world. We have sketched above a way of effecting the transition from texts which are decisionally unimodal, simple ones such as the detective story, to those which are n-modal. A work which embodies the relational paradigmatics of the "compass card" thereby sets up an undecidability about its own meaning in that it persistently defies that "instrument of semantic diagnosis" which every human head contains. There then takes place the stabilization of a shaky equilibrium at the crossroads formed by the text itself, since we cannot even say whether it is definitely in earnest or definitely ironic, whether it belongs to the one world or to the other, whether it elevates our vale of tears to the level of transcendence (as some critics said about Kafka’s Das Schloss) or whether on the contrary it degrades the beyond to the temporal plane (as others said about Das Schloss), whether it is a parable with a moral expressed by symbols from the unconscious (this is the thesis of psychoanalytic criticism), or whether it constitutes "the fantastic without limits"-which last is the dodge our structuralist uses.

It is strange that no one is willing to admit the fact of the matter: that the work brings into head-on collision a swarm of conflicting interpretations, each of which can be defended on its own grounds. If what we had before us were a logical calculus, the sum of these conflicting judgments would clearly be zero, since contradictory propositions cancel one another out. But the work is just not a logical treatise, and therefore it becomes for us, in its semantic undecidability, a fascinating riddle. "Single-axis" structuralism fails utterly for it, but the mechanism of undamped oscillation of the reader’s surmises can be formalized by a topology of multiple decision-making, which in the limit turns the compass card into a surface representing continuous aberrations of the receiver. However, the structuralist model even as we have thus amended it is not fully adequate to a work such as Kafka’s. It falls short because its axiomatic assumption of disjointness of opposed categories (allegory : poetry, irony : earnestness, natural : supernatural) is altogether false. The crux lies in the fact that the work can be placed on the natural and the supernatural level at the same time, that it can be at once earnest and ironic, and fantastic, poetic and allegorical as well. The "at the same time" predicated here implies contradictions—but what can you do, if such a text is founded just on contradictions? This is made plain by the throng of equally justified but antagonistic interpretations which battle vainly for supremacy, i.e. for uniqueness. It is only mathematics and logic and following their example mathematical linguistics that fear contradictions as the Devil fears holy water. Only these can do nothing constructive with contradictions, which put an end to all rational cognition. What is involved is a trap disastrous for epistemology, in that it is an expression which contradicts itself (much like the classic paradox of the Liar). Yet literature manages to thrive on paradoxes, if only on ones strategically placed—precisely these constitute its perfidious advantage! Not, to be sure, from its own resources. It has not invented such horrendous powers for itself. We find logical contradictions ready made, firstly in culture: for-- to take the first example to hand--according to the canons of Christianity, whatever happens happens naturally, and at the same time it happens by the will of God, since nothing can be apart from this. The non-temporal order thus coexists with the temporal-- eternity is in every moment and in every inch. The collisions of behavior provoked by this "overlapping" predication are buffered by successive interpretations of dogma, e.g. in a species of theological consent to the use of anesthesia in childbirth. Nonetheless there is a contradiction involved which culminates in "Credo, quia absurdum est." Secondly, overlapping categorizations of percepts become the norm in dreams as well as in hyponoic states, thus not only in psychiatric symptomatology (cf. Ernst Kretschmer, Medizinische Psychologie). The coexistence in apperception of states of affairs which exclude one another both empirically and logically is, consequently, a double regularity—cultural and psychological—on which structuralism finally breaks every bone in all its "axes." Thus the whole literary-critical procrustics or catalog of adulterations, errors and oversimplifications formed by this Introduction à la litérature fantastique is of value only as an object lesson illustrating the downfall of a precise conceptual apparatus outside its proper domain.

WE STILL HAVE WITH US THE DILEMMA of the hardheaded reader, who, if he is not scared by a ghost story, relabels it with respect to genre. Todorov would hold such a receiver to be an ignoramus who ought to keep his hands off literature. But when we examine the situation in which someone reads an "uncanny" or a "tragic" text and splits his sides laughing, we will realize that this situation can be explained in either of two ways. Perhaps the reader is in fact a primitive oaf who is too immature to appreciate the work, and that is an end of the problem. Or perhaps the work is kitsch and he who laughs at it is an experienced connoisseur of literature, so that he cannot take seriously what the work presents as serious, i.e. he has outgrown the work. In the second case the text really does change its genre: from a story about spirits (intentionally uncanny) or about galactic monarchs (intentionally science-fictional) or about life in high society (intentionally edifying romance) it turns into an unintentional humoresque.

Todorov bars saying anything at all about an author’s intentions—to mention these amounts to covering oneself with the disgrace of ‘fallacia intentionalis.' Structuralism is supposed to investigate texts only in their immanence. But if one is free to recognize, as Todorov does, that a text implies a reader (not as a concrete person but as a standard of reception), then in accord with a rule of symmetry one should recognize that it also implies an author. Both of these concepts are indissolubly connected with the category of messages, since a message, in information theory, must have a sender and a receiver.

The words of Roger Caillois about "the irreducible impression of strangeness" as a touchstone for the fantastic represent the psychological correlate of the linguistic state of things constituted by the full-valued character of the artistic text, which guarantees that it is not kitsch. The irreducibility of the impression certifies the authentic values of the text and thereby abolishes the relativism typical for writing with unwarranted pretensions, which produces kitsch as an incongruity between intention and realization.

The relativism of kitsch lies in the fact that it is not kitsch for all readers, and what is more it cannot be recognized as kitsch by those who esteem it. Kitsch identified as such forms a special case of paradox within the set of literary works: namely, contradiction between the reactions anticipated by the text and the reactions which its reading actually evokes. For the uncanny is incompatible with nonsense, physics with magic, the sociology of the aristocracy with the scullery’s notions about it, and the process of cognition with the adventures of puppets called scientists. Thus kitsch is a product counterfeited to pass for what it is not. The contradictions in interpretation of Kafka’s writings not only can but must be grasped by the reader; only so, thanks to "indecision of manifold scope," will he apprehend the aura of mystery established by the text. Per contra, the contradiction specific to kitsch must remain unrecognized by its readers, since otherwise generic disqualification of what has been read will take place. The reading of kitsch as kitsch is non-immanent-the reader appeals to his own superior knowledge about how a work of the given kind ought to look, and the chasm separating what ought to be from what in fact is amuses him (or offends him).

Because our superior knowledge decreases as the themes of literature become increasingly remote from reality, kitsch takes up residence in regions inaccessible to the reader. in the palace, in the far future, among the stars, in history, in exotic lands. Every literary genre has its masterwork-ceiling, and kitsch, by a tactics of crude mimicry, pretends to have soared to such an altitude. Todorov, fettered by the immanence of his procedures, has deprived himself of any possibility of recognizing mimicry of values, and accordingly his implicit reader must, by dint of solemn exertions, see to it that the silliest twaddle about spirits sends chills up and down his spine. On pain of a structuralist curse he is forbidden to poke fun at such rubbish; since structuralism establishes absolute equality in literature, the right of citizenship which the text usurps for itself is a sacred thing.

A possible rejoinder at this point would be that idiotic stories are written for idiotic readers. And indeed, we observe this state of affairs in the book market, dominated by the laws of supply and demand. But this is not an extenuating circumstance for a theory of literature. A "theory" is synonymous with a generalization which applies without exception to all elements of the set under investigation. Since the structuralists’ generalizations balk at applying thus, or, more precisely, because when they are made to apply thus everywhere they yield such nonsense as no advocate of the school would like to acknowledge (for structural equivalence democratically places the counterfeit on an equal footing with the masterpiece), the theoreticians carry out certain sleight-of-hand manipulations when they assemble their materials for public dissection. They place on their operating table, to wit, only what has already earned a respectable reputation in the history of literature, and they conjure away under the table works that are structurally of the same kinds but artistically trashy. They have to proceed thus, because their method impels them toward simple texts such as the detective story, their over-weening ambitions, on the other hand, toward celebrated works. (Kitsch, being subject to relativization in the process of reception, is not the structurally simplest case, for it seeks to be one thing and is in fact another; the detective story, on the other hand, devoid of pretensions, is decisionally unimodal.)

Now we can more readily understand the makeup of Todorov’s bibliography, as to the names (Balzac, Poe, Gogol, Hoffmann, Kafka) and the works it includes. The theoretician has taken as his "sample" that which could not involve him in difficulties, since it had already passed its cultural screening examination and by that token could give him no trouble. A therapist, if he were to proceed analogously, would take as patients only robust convalescents. A physicist would test his theory only on facts that he knew beforehand would confirm it, carefully avoiding all others. Let us spare the structuralist the description which the philosophy of science would give to such a method of selecting "representative samples."

A theory of literature either embraces all works or it is no theory. A theory of works weeded out in advance by means beyond its compass constitutes not generalization but its contrary, that is particularization. One cannot when theorizing discriminate beforehand against a certain group of works, i.e. not bring them under the scope of analysis at all. A taxonomically oriented theory can set up a hierarchy in its subject matter, i.e. assign non-uniform values to the elements of the entire set under investigation, but it should do this openly, not on the sly, and throughout its whole domain, showing what sort of criteria it employs for making distinctions and how they perform their tasks of evaluation.

These obligations are binding not for humanistic studies alone. They stem from the set of directives to which all scientific cognition is subject. A zoologist cannot ignore cockroaches because they’re such nasty little beasties, nor a cosmologist ignore the energy balance of quasars because it makes his calculations blow up in his face. The sleight-of-hand artist’s activities are not always and everywhere admirable. So, we conclude, if structuralism desires to avoid expulsion from among the sciences, it must rebuild itself completely from the ground up, since in its present state it is—in the words of Pierre Bertaux—a procedure which from its point of departure in logic has strayed into useless mythology.


1. Bertaux is a Germanist, and he published the article quoted, "Innovation als Prinzip," in German in the volume Das 198. Jahrzehnt (Christian Wegner Verlag, 1969).—SL. The passage given in German in Dr. Lem’s original text (from which the first sentence has been reduced to the bracketed phrase in our translation) reads as follows: "Unter ‘Diagonalwissenschaften’ (um den Ausdruck von Roger Caillois aufzunehmen) verstehe ich ungerfähr das, was man auch ‘formalistische’ Wissenschaften nennt, also Disziplinen, deren Gebiet sich quer durch die herkömmlichen Fächer der Realwissenschaften zieht.... Eine Zeitlang hat man hoffen können, der Ansatz zu einer ähnlichen Formalisierung der Humanwissenschaften sei vom Strukturalismus zu erwarten. Leider sieht es heute aus, also ab gerade die lautesten Vertreter des Strukturalismus ihn zu einer Mythologie hätten entarten lassen-und nicht einmal zu einer brauchbaren. Das Gerede, das jetzt den Namen Strukturalismus trägt, hat den ursprünglich in ihm enthaltenen wissenschaftlichen Ansatz wahrscheinlich tödlich getroffen." -CN, RDM, DS.

2. Translated by Richard Howard (Cleveland/London: The Press of Case Western Reserve University 1973) from Introduction à la littérature fantastique (Editions du Seuil 1970). All quotations from Todorov are from the pages of this translation. —RDM.



 Since structuralism in literary studies is largely of French origin, this attempt to ruin its reputation takes as its motto the words of a Frenchman, Pierre Bertaux: "At one time it was hoped that the beginnings of a formalization of the humanities analogous [to that of the sciences] could be expected from structuralism. Unfortunately, it appears today that precisely the loudest advocates of structuralism have let it degenerate into a mythology—and not even a useful one." I fully agree with this verdict. However, inasmuch as it is difficult to expose in a single article the barrenness of a whole school of thought—one moreover which has spawned divergent tendencies, since every author has his own "vision" of the subject—I will limit myself to dissecting Tzvetan Todorov’s book The Fantastic. The author begins by deriding the investigator who would, before proceeding to description of a genre, engage in endless reading of actual works. Todorov’s "sample" of works discussed, as displayed in his bibliography, is astonishing. Among its twenty-seven titles we find no Borges, no Verne, no Wells, nothing from modern fantasy: all of SF is represented by two short stories. We get, instead, E.T.A. Hoffman, Potocki, Balzac, Poe, Gogol, Kafka—and that is about all. What this structural account proclaims to us as the bounds of the fantastic is really quite an antique piece of furniture: the bed of Procrustes. 

moonbut.gif (4466 bytes)Back to Home