Science Fiction Studies

# 6 = Volume 2, Part 2 = July 1975

On Lem on Todorov


As an admirer of Stanislaw Lem—both as a speculative fabulator and as a poetician—I was grieved by his attack on Tzvetan Todorov in the essay "Todorov's Fantastic Theory of Literature" (SFS 1:227-37). That I have a personal stake in this must be admitted. I have written an introduction for the Cornell Press's new paperback version of the very book Lem attacked so virulently, and I am presently translating Todorov's Poétique de la prose. Also, I have written with some enthusiasm—though tempered, I trust, with critical restraint—of structuralism itself, and structuralism was a major object of Lem's polemic. I make this clear so that the reader may know my bias, but at the same time, I will try to keep that bias under control in the interest of objectivity and that truth to which we all aspire but never attain.

Polemical criticism has its uses. Certainly it can shake us out of critical complacency. A polemical review of Todorov's book will appear soon in Novel, a magazine of which I am the book review editor. But there are polemics and polemics. It seems to me that a polemic is valuable to the extent that it challenges ideas, and that it loses in value to the extent that it seems aimed at inflicting psychological damage on a human adversary. And Lem is plainly guilty of this. It is just a style of debate, you say, but it is an unfortunate style, and unworthy of Lem, whose words are weighty enough simply because they have his extraordinary achievements behind them.

When Lem turns to ideas, he says some very shrewd things about the limitations of certain kinds of structural criticism, though I think he is wrong in postulating a monolithic "structuralism" to which all of his criticisms apply. His criticism of Todorov, in fact, takes a familiar pattern, which may be distinguished as a generic feature of polemical criticism as a whole. First he labels his antagonist: in this case, as a "structuralist." Then he defines "structuralism" in a very limited way. And finally he abuses the "structuralist" for abandoning "structuralism" whenever he fails to conform to the narrow definition provided. This is a classic double-bind situation, typical of psychological in-fighting but not very helpful in critical discussion.

But rather than attack Lem, and generate a new polemic of my own, or simply defend Todorov, I will try to mediate here if I can. It seems to me that the main point at issue between Todorov and Lem involves the number and variety of texts that are going to be called "fantastic." Todorov wants to be exclusive. Lem wants to be inclusive. (But we should remember that only a few issues back, in the first volume of SFS, in Lem's own "structural" approach to the genre, he wanted to exclude most so-called sf from his privileged category of "realism about some other place on the space-time continuum," presumably dumping the left-overs into a catchall category of fantasy.) According to Todorov, the whole point of generic analysis is to refine generic awareness by excluding, sharpening critical focus until a closely related group of texts may be studied together. Lem, when he is being "structural" about sf, seems to work the same way, but he is quite content to include as much as possible under the heading of fantasy. Thus it is not at all clear how he would go about distinguishing between "fantastic theology," for instance, and "real theology." Not to mention the possible confusions between the fantastic theology of the present and the real theology of some other place on the space-time continuum.

But I don't want to turn this into a polemic against Lem's generic criticism, which I actually find stimulating and fruitful. The point is simply that generic criticism is very difficult, and there is no easy middle ground in it between defining too precisely and not precisely enough. Such criticism also poses immense terminological problems, forcing the critic either to adapt old terms to new purposes or to invent new ones, neither of which is always a happy solution. The essential conflict between Lem and Todorov, as I see it, lies in this area of terminology, specifically in the word "fantasy" itself. Todorov has taken, here, a word normally used to designate a large and spongy tract of literature and given that name to a narrow pathway. From this, much disputation has begun to spring. Yet I have no doubt that the genre Todorov had studied in his book is as real as any other generic grouping. Nor does Lem seem to doubt its existence. On the other hand, it is clear to me that this genre is only a small part of what we usually call "fantasy." Todorov calls the larger territory simply "the imaginary," and he locates his "fantastic" on the interface between the real and the imaginary just as Lem locates his "real" on the interface between myth and fairy tale). If Todorov had called his intermediate genre the "uncanny," or given it some other less broadly designative term, much polemicizing might have been avoided. As a mediator, here, I am perfectly willing to say that I wish he had not tried to take a term that usually designates the radically unreal and make this same term designate a hesitation between the real and unreal. In ordinary English usage, at any rate, "uncanny" is much closer to the mark. Here Todorov's English translator has not helped much, by translating étrange as uncanny. Still, if we could separate a dispute about names from the dispute about concepts, we might find that there is actually less to dispute about than we had supposed.

There is much in Lem's more general attack on structuralism that I should like to discuss, but SFS is probably not the place to do this, especially with the space limitations the editors have set for this response to Lem's critique. Here, I must content myself with urging readers to look into Todorov for themselves, since Lem has not represented his adversary's position as clearly and fairly as he might have.



It is difficult for me to imagine that anyone familiar with Tzvetan Todorov's Introduction à la littérature fantastique (misleadingly abbreviated to simply The Fantastic for the English translation) could take seriously Stanislaw Lem's misreading of it in SFS #4. However, for the benefit of those as yet unexposed to Todorov and to structuralism, some sort of response seems desirable, particularly in light of Lem's otherwise well-deserved prestige in these pages. (Note: in what follows, not having seen the English-language edition, I will rely exclusively on the French text published in Paris in 1970, which is also, I understand, the text used by Lem.)

Todorov's purpose in that portion of the book that Lem attacks is a simple one: to investigate a literary category, a "genre," characterized by a particular effect, and to discover the rule that defines this category. It is Todorov's fate, or perhaps his carelessness (not seeing Lem in the bushes, waiting to pounce), to have chosen for the name of this effect a word, "fantastic," which means something quite different to Lem. It is, in turn, Lem's fate not to have noticed this difference in terminology, and to have become hopelessly confused on almost every point he takes up.

The "fantastic," according to Todorov, is characterized by the hesitation of a being (the hero of the story and the reader who "identifies" with this hero) who knows only natural laws when faced with an event in appearance supernatural. This hesitation finds the hero suspended between abandoning her or his dependence on rational explanation on the one hand, and, on the other, constructing a complicated and unlikely but rationally possible explanation for the apparently supernatural events.

Historically speaking, prior to what we refer to as the "Enlightenment," there could be no such hesitation. The supernatural was accepted as a part of life. Witches and God co-existed with men and women, and a story could, in Todorov's terms, be "marvelous," but never "fantastic." Examples abound: Sinbad the Sailor, fairy tales, chivalric romances. At the other end—our end—of the nineteenth century, with the psychoanalytic discovery of the unconscious, there is again no hesitation. The witness to bizarre events, or at least the reader of the story, knows them to be the creations of his or her own mind. A story then may be "strange" (étrange, inexplicably translated as "uncanny" by Richard Howard), but, again, never "fantastic," science fiction and Todorov's careless remarks about it notwithstanding. For Todorov, science-fiction is a species of the marvelous, but the sense in which "robots, extraterrestrial beings, the whole interplanetary context" are supernatural is entirely different. Here the marvelous and the strange intersect without creating that cognitive hesitation characteristic of the fantastic, for the explanation of the events, while currently impossible (we as yet know no interplanetary beings) is implicitly rational (we recognize the possibility that we will know such beings in another time).

Thus the hesitation of the reader-hero in the fantastic reflects the hesitation of history, and Todorov's fantastic is an historical form, nothing but the backside of nineteenth century positivism, with its foreclosure of the supernatural ("la littérature fantastique n'est rien d'autre que la mauvaise conscience de ce dix-neuvième siècle positiviste" (p176). Should this conclusion hold up—and Lem, at least, does not attack it—the value of Todorov's book seems to be assured, whatever one thinks of his methodology.

Lem misunderstands not only the object but also the theoretical framework of Lem's study, at at least two points: the notion of "genre" and the nature of "Todorov's axis."

Historical genres do not correspond to biological species, despite Todorov's attempt to rehabilitate Propp's morphological analogy. There is, as Lem points out, no such thing as a "normal" story in the sense that there is a "normal" (i.e., "abstract") tiger. However, there is no such thing as a "normal" Victorian house, either. Yet one need not look at them all to discuss the "historical genre" of Victorian houses.

Of course, this notion of genre is restrictive, and "only mass literature must call in the notion of genre" (translated from pp10-11 of the French text). (Lem makes the same point, remarking that "the more inferior and paradigmatically petrified the texts which it undertakes to anatomize, the more readily [Todorov's method] will reveal structures," but careless continues, "Todorov, not surprisingly, omits to draw this conclusion" [SFS 1:228].) But since all works inevitably find their place somewhere between generic conventions and the breaking of them (one need only think of Solaris and the conventions of the alien-contact novel), the notion of genre is more or less applicable to all texts.

Todorov, however, is not speaking directly of historical genres, but of "theoretically possible" genres, elementary and complex, defined by the presence or absence of a single structural trait or a conjunction of such traits, respectively. These genres are not, as Lem seems to understand, mutually exclusive categories into which texts are pigeon-holed. A single text may belong to any number of theoretically possible genres, or a given genre may be represented by no texts. The usefulness of Todorov's method, in its movement from this level of theory back to practice lies in the adequacy of the following observation: "On all evidence, the historical genres are a subset of the set of complex theoretically possible genres" (translated from p35 of French text).

Lem's chief theoretical (as opposed to bilious) objection to Todorov's method lies elsewhere:

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 the 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. (SFS 1:232) 

This takes the form of a general protest against structuralism on a basic theoretical level. Sticking to the example before us, it is clear that if a work is characterized by the presence or absence of a single structural trait, it either belongs or does not belong to the genre defined with respect to that trait. With a complex genre, defined by a conjunction of traits, the question of partial membership can of course be raised, but it would seem obvious that the concept of  membership is too crude, too simple to deal with it. One detects here a whiff of that crude desire for quantification characteristic of "nineteenth century positivism."

However, Lem here seems to be talking about two things at once. The Todorovian axis referred to stretches from the "marvelous" to the, "strange," and the "fantastic" is precisely the "excluded middle," defined by the balance, the cognitive hesitation, between two traits: the supernatural and the complex natural explanations for events, respectively. Here, at least, Lem's apparent desire to place works along a continuum would seem to meet with some satisfaction.



My brief response comes tripartite: (1) preliminaries, (2) evidence against the Todorovian doctrine, and (3) general remarks.

1. I concede that to show the deficiency of Todorov's conception is not to refute all existing schools of structuralism. I concede too that there are in Todorov's book some valid observations on the early period of a subclass of the "Fantastic" in the 19th century, but still maintain that he failed to hit the important target at which he aimed, his generalizations being substantially indefensible and misrepresentative of the real complexity of the generic problems at stake.

2. Let us imagine a novel dealing with the conflict of opinions arising from an event like the following. There is preserved in a Roman church in a phial the blood of a sainted martyr, and this blood, when shown to a large crowd on a certain day of the year, becomes liquid. The central dilemma of the novel is whether this liquefaction of the blood in the glass container occurs naturally (e.g., catalyzed by sunlight) or whether it represents a "true miracle." Mr. Astle writes, "The 'fantastic,' according to Todorov, is characterised by the hesitation of a being (the hero of the story and the reader who 'identifies' with this hero) who knows only natural laws when faced with an event in appearance supernatural." If so, then the novel described above must per definitionem be classified as "fantastic." But this would contradict what any sane reader would say—that the novel is a realistic novel, for the facts described in it are real, and the problematics involved are of a religious instead of a fantastic character. Or take another novel, one concerned with the central controversy of anthropogenesis: what caused the change of some hominids into men—natural evolution only, or some divine intervention? I hope that no one will dare to say that this question is of a fantastic character in the meaning that Todorov has given the term. Nevertheless we have before us in both cases the typical hesitation (according to his definition) between natural and supernatural laws. If we follow Todorov further (as good believers should), we will be forced to conclude that all head-on collisions between the "sacred" and "profane" mainstreams of human thought are coextensive with the "fantastic." Every literary hero who as a "doubting atheist" or "doubting transcendentalist" hesitates between naturalistic and super-naturalistic ontology becomes, ipso facto, according to the Todorovian doctrine, the hero of a fantastic story. But this is already reductio ad absurdum of the whole concept, is it not?

My next example is the "sealed room" subset of crime stories. When the impenetrability of the sealed room to all "natural means" has been described in details of the utmost completeness, the evidence thus accumulated points to the fact that the crime could not have been accomplished "naturally." If so, then every reader should hesitate between the Natural and Supernatural explanations, for the situation embodies perfectly the structure of Todorov's "fantastic." But there is no trace of such hesitation in the readers. No fan of crime stories will anticipate a "supernatural explanation" of the crime. Why not? The answer is simplicity in itself: the readers are accustomed to the rules of this play, and its first rule denies the author any use of "accomplices from heaven or hell." So if a story should end with the Detective's assertion that "this crime was perpetrated by some Superhuman Agent," the readers would tell the author a cheat. This shows that the reader does know the genre of the text beforehand, and because of this the reading is never of an "immanent" character as structuralism would have it. Todorov has himself perpetrated petitio principii, giving the explanandum for the explanans.

In my SFS essay I demonstrated that there are literary works, as those of Borges, that do not comply with the Todorovian conditions of "fantastic" but are nevertheless appreciated as "fantastic" by the average reader, and above I have shown that there are literary works that comply perfectly with all the conditions necessary and sufficient, according to Todorov, to produce the "fantastic effect" but that have nothing in common with our intuitions of what we are accustomed to regarding as "fantastic." What follows can now be said easily. Sometimes the Todorovian conditions are fulfilled by what we are accustomed to calling "fantastic," and sometimes they are not. But a theory sometimes confirmed and sometimes confuted is no valid theory at all.

3. I do not imagine, however, that I have convinced my critics with the arguments made above, so now I come to the last part of my counterargument: the origin of the principal differences between our attitudes in the field of cognitive generalizations. I have had no formal training in the humanities; in this field I am an intruder from the realm of natural science, where I learned what a theory is. It is a generalization from a univocally delimited set of facts falsifiable by means of experiment. Please note that neither of my critics is defending the theory of the fantastic as it was stated by Todorov; each is instead defending his own version of that theory. Should more parties come into the field, then there would begin a further disintegration of Todorovian doctrine, a divergent multiplication of theoretical attitudes without end. This situation I find intolerable. One must have established means to prove or refute any given theory in the humanities as well as in physics. I am well aware of the monstrous difficulties that arise when one attempts to introduce into the theory of literature this kind of methodological rigor as standard. I have myself written an opus, "Philosophy of Chance," with the subtitle, "Literature in the Light of Empiricism" (Cracow: Wyd. Literackie, 2nd edn, 1975), in which in the first part of the second volume I criticize various schools of structuralism. This work I cannot summarize here. I can only describe the context of my grievance against structuralism.

We are now in the midst of an information explosion, and one of its centers is in the arts, where the danger comes from what I will call the limits to cultural growth. To transgress these limits is to release the entropy of culture in an uncontrolled way. This growth, accelerating, produces disorder (= entropy) and thus disrupts the self-sustaining chain-processes of cultural evolution; or, to put it another way, the rising chaos incapacitates all factors responsible for the value-oriented "natural selection" of paradigms in culture. This "natural selection" has given to the culture of the past a homeostatic equilibrium between anachronization and modernization of creativity. But today the traditional values are invalidated and the new ones are born immature and with a constantly decreasing life span. Any evolutionary process in which there is no adequacy of the mutation and stabilization rates, in which the stabilizing factors are held down while the mutational ones multiply exponentially, must collapse. Confronted with the premonitory symptoms of such collapse, what we need most in the theoretical field are, first, generalizations emphasizing the diachronical linkage of culture, and, second, generalizations capable of axiological differentiation. Both these tasks are neglected by structuralism. It is neither diachronical in its typical proceedings, nor axiologically selective. It is hermeneutical proliferation at its worst, since it suffices to know some logic (some second-rate logic, I would add) and to have a McLuhanian imagination to produce each day a lot of new structuralistic hypotheses, promoting the already great confusion of concepts. Because of this, structuralism is not a neutral constituent of our intellectual life, but a paralyzing agent, the modern reincarnation of scholasticism, even if some of its fruits are not as toxic as Todorov's Introduction à la littérature fantastique.

Back to Home