Science Fiction Studies

#1 = Vol 1, Part 1 = Spring 1973

Stanislaw Lem

On the Structural Analysis of Science Fiction

In the early stages of literary development the different branches of literature, the genological types, are distinguished clearly and unmistakably. Only in the more advanced stages do we find hybridization. But since some crossbreedings are always forbidden, there exists a main law of literature that could be called incest prohibition; that is, the taboo of genological incest.

A literary work considered as a game has to be played out to the finish under the same rules with which it was begun. A game can be empty or meaningful. An empty game has only inner semantics, for it derives entirely from the relationships that obtain between the objects with which it is played. On a chessboard, for example, the king has its specific meanings within the rules of the play, but has no reference outside the rules; i.e., it is nothing at all in relation to the world outside the confines of the chessboard. Literary games can never have so great a degree of semantic vacuum, for they are played with "natural language", which always has meanings oriented toward the world of real objects. Only with a language especially constructed to have no outward semantics, such as mathematics, is it possible to play empty games.

In any literary game there are rules of two kinds: those that realize outer semantic functions as the game unfolds and those that make the unfolding possible. "Fantastic" rules of the second kind--those that make the unfolding possible--are not necessarily felt as such even when they imply events that could not possibly occur in the real world. For example, the thoughts of a dying man are often detailed in quite realistic fiction even though it is impossible, therefore fantastic, to read the thoughts of a dying man out of his head and reproduce them in language. In such cases we simply have a convention, a tacit agreement between writer and reader-in a word, the specific rule of literary games that allows the use of nonrealistic means (e.g., thought-reading) for the presentation of realistic happenings.

Literary games are complicated by the fact that the rules that realize outer semantic functions can be oriented in several directions. The main types of literary creation imply different ontologies. But you would be quite mistaken if you believed, for example, that the classical fairy tale has only its autonomous inner meanings and no relationship with the real world. If the real world did not exist, fairy tales would have no meaning. The events that occur in a myth or fairy tale are always semantically connected with what fate has decreed for the inhabitants of the depicted world, which means that the world of a myth or fairy tale is ontologically either inimical or friendly toward its inhabitants, never neutral; it is thus ontologically different from the real world, which may be here defined as consisting of a variety of objects and processes that lack intention, that have no meaning, no message, that wish us neither well nor ill, that are just there. The worlds of myth or fairy tale have been built either as traps or as happiness-giving universes. If a world without intention did not exist; that is, if the real world did not exist, it would be impossible for us to perceive the differentia specifica, the uniqueness, of the myth and fairy-tale worlds.

Literary works can have several semantic relationships at the same time. For fairy tales the inner meaning is derived from the contrast with the ontological properties of the real world, but for anti-fairy tales, such as those by Mark Twain in which the worst children live happily and only the good and well-bred end fatally, the meaning is arrived at by turning the paradigm of the classical fairy tale upside down. In other words, the first referent of a semantic relationship need not be the real world but may instead be the typology of a well-known class of literary games. The rules of the basic game can be inverted, as they are in Mark Twain, and thus is created a new generation, a new set of rules--and a new kind of literary work.

In the 20th century the evolution of mainstream literary rules has both allowed the author new liberties and simultaneously subjected him to new restrictions. This evolution is antinomical, as it were. In earlier times the author was permitted to claim all the attributes of God: nothing that concerned his hero could be hidden from him. But such rules had already lost their validity with Dostoyevsky, and god-like omniscience with respect to the world he has created is now forbidden the author. The new restrictions are realistic in that as human beings we act only on the basis of incomplete information. The author is now one of us; he is not allowed to play God. At the same time, however, he is allowed to create inner worlds that need not necessarily be similar to the real world, but can instead show different kinds of deviation from it.

These new deviations are very important to the contemporary author. The worlds of myth and fairy tale also deviate from the real world, but individual authors do not invent the ways in which they do so: in writing a fairy tale you must accept certain axioms you haven't invented, or you won't write a fairy tale. In mainstream literature, however, you are now allowed to attribute pseudo-ontological qualities of your personal, private invention to the world you describe. Since all deviations of the described world from the real world necessarily have a meaning, the sum of all such deviations is (or should be) a coherent strategy or semantic intention.

Therefore we have two kinds of literary fantasy: "final" fantasy as in fairy tales and SF, and "passing" fantasy as in Kafka. In an SF story the presence of intelligent dinosaurs does not usually signal the presence of hidden meaning. The dinosaurs are instead meant to be admired as we would admire a giraffe in a zoological garden; that is, they are intended not as parts of an expressive semantic system but only as parts of the empirical world. In "The Metamorphosis", on the other hand, it is not intended that we should accept the transformation of human being into bug simply as a fantastic marvel but rather that we should pass on to the recognition that Kafka has with objects and their deformations depicted a socio-psychological situation. Only the outer shell of this world is formed by the strange phenomena; the inner core has a solid non-fantastic meaning. Thus a story can depict the world as it is, or interpret the world (attribute values to it, judge it, call it names, laugh at it, etc.), or, in most cases, do both things at the same time.

If the depicted world is oriented positively toward man, it is the world of the classical fairy tale, in which physics is controlled by morality, for in a fairy tale there can be no physical accidents that result in anyone's death, no irreparable damage to the positive hero. If it is oriented negatively, it is the world of myth ("Do what you will, you'll still become guilty of killing your father and committing incest."). If it is neutral, it is the real world--the world which realism describes in its contemporary shape and which SF tries to describe at other points on the space-time continuum.

For it is the premise of SF that anything shown shall in principle be interpretable empirically and rationally. In SF there can be no inexplicable marvels, no transcendences, no devils or demons--and the pattern of occurrences must be verisimilar.

And now we come near the rub, for what is meant by a verisimilar pattern of occurrences? SF authors try to blackmail us by calling upon the omnipotence of science and the infinity of the cosmos as a continuum. "Anything can happen" and therefore "anything that happens to occur to us" can be presented in SF.

But it is not true, even in a purely mathematical sense, that anything can happen, for there are infinities of quite different powers. But let us leave mathematics alone. SF can be either "real SF" or "Pseudo-SF".

When it produces fantasy of the Kafka kind it is only pseudo-SF, for then it concentrates on the content to be signaled. What meaningful and total relationships obtain between the telegram "mother died, funeral Monday" and the structure and function of the telegraphic apparatus? None. The apparatus merely enables us to transmit the message, which is also the case with semantically dense objects of a fantastic nature, such as the metamorphosis of man into bug, that nevertheless transmit a realistic communication.

If we were to change railway signals so that they ordered the stopping of trains in moments of danger not by blinking red lights but by pointing with stuffed dragons, we would be using fantastic objects as signals, but, those objects would still have a real, non-fantastic function. The fact that there are no dragons has no relationship to the real purpose or method of signaling.

As in life we can solve real problems with the help of images of nonexistent beings, so in literature can we signal the existence of real problems with the help of prima facie impossible occurrences or objects. Even when the happenings it describes are totally impossible, an SF work may still point out meaningful, indeed rational, problems. For example, the social, psychological, political, and economic problems of space travel may be depicted quite realistically in SF even though the technological parameters of the spaceships described are quite fantastic in the sense that it will for all eternity be impossible to build a spaceship with such parameters.

But what if everything in an SF work is fantastic? What if not only the objects but also the problems have no chance of ever being realized, as when impossible time-travel machines are used to point out impossible time-travel paradoxes? In such cases SF is playing an empty game.

Since empty games have no hidden meaning, since they represent nothing and predict nothing, they have no relationship at all to the real world and can therefore please us only as logical puzzles, as paradoxes, as intellectual acrobatics. Their value is autonomous, for they lack all semantic reference; therefore they are worthwhile or worthless only as games. But how do we evaluate empty games? Simply by their formal qualities. They must contain a multitude of rules; they must be elegant, strict, witty, precise, and original. They must therefore show at least a minimum of complexity and an inner coherence; that is, it must be forbidden to make during the play any change in the rules that would make the play easier.

Nevertheless, 90 to 98 percent of the empty games in SF are very primitive, very naive one-parameter processes. They are almost always based on only one or two rules, -and in most cases it is the rule of inversion that becomes their method of creation. To write such a story you invert the members of a pair of linked concepts. For example, we think the human body quite beautiful, but in the eyes of an extraterrestrial we are all monsters: in Sheckley's "All the Things You Are" the odor of human beings is poisonous for extraterrestrials, and when they touch the skin of humans they get blisters, etc. What appears normal to us is abnormal to others--about half of Sheckley's stories are built on this principle. The simplest kind of inversion is a chance mistake. Such mistakes are great favorites in SF: something that doesn't belong in our time arrives here accidentally (a wrong time-mailing), etc.

Inversions are interesting only when the change is in a basic property of the world. Time-travel stories originated in that way: time, which is irreversible, acquired a reversible character. On the other hand, any inversion of a local kind is primitive (on Earth humans are the highest biological species, on another planet humans are the cattle of intelligent dinosaurs; we consist of albumen, the aliens of silicon; etc.). Only a non-local inversion can have interesting consequences: we use language as an instrument of communication; any instrument can in principle be used for the good or bad of its inventor. Therefore the idea that language can be used as an instrument of enslavement, as in Delany's Babel-17, is interesting as an extension of the hypothesis that world view and conceptual apparatus are interdependent; i.e., because of the ontological character of the inversion.

The pregnancy of a virgo immaculata; the running of 100 meters in 0.1 seconds; the equation 2 x 2-7; the pan-psychism of all cosmic phenomena postulated by Stapledon: these are four kinds of fantastic condition.

1. It is in principle possible, even empirically possible, to start embryogenesis in a virgin's egg; although empirically improbable today, this condition may acquire an empirical character in the future.

2. It will always be impossible for a man to run 100 meters in 0.1 seconds. For such a feat a man's body would have to be so totally reconstructed that he would no longer be a man of flesh and blood. Therefore a story based on the premise that a human being as a human being could run so fast would be a work of fantasy, not SF.

3. The product of 2 x 2 can never become 7. To generalize, it is impossible to realize any kind of logical impossibility. For example, it is logically impossible to give a logical proof for the existence or nonexistence of a god. It follows that any imaginative literature based on such a postulate is fantasy, not SF.

4. The pan-psychism of Stapledon is an ontological hypothesis. It can never be proved in the scientific sense: any transcendence that can be proved experimentally ceases to be a transcendence, for transcendence is by definition empirically unprovable. God reduced to empiricism is no longer God; the frontier between faith and knowledge can therefore never be annulled.

But when any of these conditions, or any condition of the same order, is described not in order to postulate its real existence, but only in order to interpret some content of a semantic character by means of such a condition used as a signal-object, then all such classificatory arguments lose their power.

What therefore is basically wrong in SF is the abolition of differences that have a categorical character: the passing off of myths and fairy tales for quasi-scientific hypotheses or their consequences, and of the wishful dream or horror story as prediction; the postulation of the incommensurable as commensurable; the depiction of the accomplishment of possible tasks with means that have no empirical character; the pretense that insoluble problems (such as those of a logical typus) are soluble.

But why should we deem such procedures wrong when once upon a time myths, fairy tales, sagas, fables were highly valued as keys to all cosmic locks? It is the spirit of the times. When there is no cure for cancer, magic has the same value as chemistry: the two are wholly equal in that both are wholly worthless. But if there arises a realistic expectation of achieving a victory over cancer, at that moment the equality will dissolve, and the possible and workable will be separated from the impossible and unworkable. It is only when the existence of a rational science permits us to rule the phenomena in question that we can differentiate between wishful thinking and reality. When there is no source for such knowledge, all hypotheses, myths, and dreams are equal; but when such knowledge begins to accumulate, it is not interchangeable with anything else, for it involves not just isolated phenomena but the whole structure of reality. When you can only dream of space travel, it makes no difference what you use as technique: sailing ships, balloons, flying carpets or flying saucers. But when space travel becomes fact, you can no longer choose what pleases you rather than real methods.

The emergence of such necessities and restrictions often goes unnoticed in SF. If scientific facts are not simplified to the point where they lose all validity, they are put into worlds categorically, ontologically different from the real world. Since SF portrays the future or the extraterrestrial, the worlds of SF necessarily deviate from the real world, and the ways in which they deviate are the core and meaning of the SF creation. But what we usually find is not what may happen tomorrow but the forever impossible, not the real but the fairy-tale-like. The difference between the real world and the fantastic world arises stochastically, gradually, step by step. It is the same kind of process as that which turns a head full of hair into a bald head: if you lose a hundred, even a thousand hairs, you will not be bald; but when does balding begin--with the loss of 10,000 hairs or 10,950?

Since there are no humans that typify the total ideal average, the paradox of the balding head exists also in realistic fiction, but there at least we have a guide, an apparatus in our head that enables us to separate the likely from the unlikely. We lose this guide when reading portrayals of the future or of galactic empires. SF profits from this paralysis of the reader's critical apparatus, for when it simplifies physical, psychological, social, economic, or anthropological occurrences, the falsifications thus produced are not immediately and unmistakably recognized as such. During the reading one feels instead a general disturbance; one is dissatisfied; but because one doesn't know how it should have been done, is often unable to formulate a clear and pointed criticism.

For if SF is something more than just fairy-tale fiction, it has the right to neglect the fairy-tale world and its rules. It is also not realism, and therefore has the right to neglect the methods of realistic description. Its genological indefiniteness facilitates its existence, for it is supposedly not subject to the whole range of the criteria by which literary works are normally judged. It is not allegorical; but then it says that allegory is not its task: SF and Kafka are two quite different fields of creation. It is not realistic, but then it is not a part of realistic literature. The future? How often have SF authors disclaimed any intention of making predictions! Finally, it is called the Myth of the 21st Century. But the ontological character of myth is anti-empirical, and though a technological civilization may have its myths, it cannot itself embody a myth. For myth is an interpretation, a comparatio, an explication, and first you must have the object that is to be explicated. SF lives in but strives to emerge from this antinomical state of being.

A quite general symptom of the sickness in SF can he found by comparing the spirit in ordinary literary circles to that in SF circles. In the literature of the contemporary scene there is today uncertainty, distrust of all traditional narrative techniques, dissatisfaction with newly created work, general unrest that finds expression in ever new attempts and experiments; in SF, on the other hand, there is general satisfaction, contentedness, pride; and the results of such comparisons must give us some food for thought.

I believe that the existence and continuation of the great and radical changes effected in all fields of life by technological progress will lead SF into a crisis which is perhaps already beginning. It becomes more and more apparent that the narrative structures of SF deviate more and more from all real processes, having been used again and again since they were first introduced and having thus become frozen, fossilized paradigms. SF involves the art of putting hypothetical premises into the very complicated stream of socio-psychological occurrences. Although this art once had its master in H.G. Wells, it has been forgotten and is now lost. But it can be learned again.

The quarrel between the orthodox and heterodox parts of the SF fraternity is regrettably sterile, and it is to be ft-award that it will remain so, for the readers that could in principle be gained for a new, better, more complex SF, could be won only from the ranks of the readers of mainstream literature, not from the ranks of the fans. For I do not believe that it would be possible to read this hypothetical, non-existent, and phenomenally good SF if you had not first read all the best and most complex works of world literature with joy (that is, without having been forced to read them). The revolutionary improvement of SF is therefore always endangered by the desertion of large masses of readers. And if neither authors nor readers wish such an event, the likelihood of a positive change in the field during the coming years must be considered as very small, as, indeed, almost zero. For it would then be a phenomenon of the kind called in futurology "the changing of a complex trend", and such changes do not occur unless there are powerful factors arising out of the environment rather than out of the will and determination of a few individuals.

POSTSCRIPT: Even the best SF novels tend to show, in the development of the plot, variations in credibility greater than those to be found even in mediocre novels of other kinds. Although events impossible from an objective-empirical standpoint (such as a man springing over a wall seven meters high or a woman giving birth in two instead of nine months) do not appear in non-SF novels, events equally impossible from a speculative standpoint (such as the totally unnecessary end-game in Disch's Camp Concentration) appear frequently in SF. To be sure, separating the unlikely from the likely (finding in the street a diamond the size of your fist as opposed to finding a lost hat) is much simpler when your standard of comparison is everyday things than it is when you are concerned with the consequences of fictive hypotheses. But though separating the likely from the unlikely in SF is difficult, it can be mastered. The art can be learned and taught. But since the lack of selective filters is accompanied by a corresponding lack in reader-evaluations, there are no pressures on authors for such an optimization of SF.


Translated by Franz Rottensteiner and Bruce R. Gillespie, with some editing by DS and RDM.



If SF is something more than fairy tale fiction, it has the right to neglect the fairy tale world and its rules. It is also not realism and has the right to neglect the methods of realistic description. Its generic indefiniteness facilitates its existence, for it is supposedly not subject to the whole range of criteria by which literary works normally are judged. SF is not allegorical, but then it says allegory is not its task: SF and Kafka are quite different. It is not realistic, but then it is not a part of realistic literature. The future? How often have SF authors disclaimed any intention of making predictions! Finally, it is the Myth of the 21st Century. But the ontological character of myth is anti-empirical, and though a technological civilization may have its myths, it cannot itself embody a myth, for myth is an interpretation, an explication, and you must have the object that is to be explicated. SF lives in but strives to emerge from this antinomical state of being. It becomes more and more apparent that its narrative structures deviate more and more from any real processes, having been used again and again since they were first introduced and having thus become frozen, fossilized paradigms. SF involves the art of putting hypothetical premises into the very complicated stream of socio-psychological occurrences. Although this art once had its master in H.G. Wells, it has been forgotten and is now lost. But it can be learned again.

moonbut.gif (4466 bytes)Back to Home