Critical Theory and Science Fiction. Carl Freedman

standards of rationality, Dante offers plausible scientific speculation as to the geography of hell in relation to that of earth (and purgatory), and that Milton does the same with regard to the substance of which angels are supposed to be made. On this level, indeed, one might even argue that Dante and Milton, in the active interest they took in the scientific developments of their own times and places, are considerably more akin to Isaac Asimov and Arthur C. Clarke than to Wordsworth and T. S. Eliot. The larger point, however, is that many of the major literary values for which science fiction is generally read are very much at work in Dante’s and Milton’s efforts to take the reader far beyond the boundaries of his or her own mundane environment, into strange, awe-inspiring realms thought to be in fact unknown, or at least largely unknown, but not in principle unknowable. It is in this sense of creating rich, complex, but not ultimately fantastic alternative worlds that Dante and Milton can be said to write science fiction. The matter can be put the other way around, as it were, by suggesting that if one were to seek, in older literature, qualities similar to those found in the multisecular historic sweep of Asimov’s Foundation (1951–1953) series or the cosmic awe at the conclusion of Clarke’s Childhood’s End (1954), one would probably do far better to go to Dante and Milton than to Romantic or post-Romantic verse, or to the realistic novel. It would seem, then, justifiable to accept the classification of Paradise Lost (1667) and Inferno (c. 1315) as science fiction.

      It would not, however, be difficult to make similar arguments with regard to a great many other texts that do not arrive in the bookshop with the rubric of “science fiction” printed on the dust jackets or back covers. The very ease with which the broadest construction of science fiction can be justified may itself arouse suspicion. As we argue that the qualities that govern texts universally agreed to be science fiction can be found to govern other texts as well, it may be difficult to see just where the argument will stop. It may even begin to appear that ultimately nearly all fiction—perhaps even including realism itself—will be found to be science fiction. Does not that conclusion preclude success in defining science fiction as a recognizable kind of fiction? In fact, I do believe that all fiction is, in a sense, science fiction. It is even salutary, I think, sometimes to put the matter in more deliberately provocative, paradoxical form, and to maintain that fiction is a subcategory of science fiction rather than the other way around. Nonetheless, the capacity of such formulations to illuminate depends upon a more conceptually specific notion of science fiction than we have suggested thus far. Merely descriptive concepts have proved adequate to expanding the term beyond the narrow pulp-centered notion; having failed to limit the category of science fiction by descriptive means, however, we are now in urgent need of a genuinely critical, analytic, definitional principle.

      By far the most helpful such principle yet suggested is that of Darko Suvin. Science fiction, he defines, is “a literary genre whose necessary and sufficient conditions are the presence and interaction of estrangement and cognition, and whose main formal device is an imaginative framework alternative to the author’s empirical environment” (emphasis deleted). He goes on to add that estrangement “differentiates [science fiction] from the ‘realistic’ literary mainstream,” while cognition differentiates it from myth, the folk tale, and fantasy.¹⁹ In this understanding, then—though Suvin does not put the matter in exactly this way—science fiction is determined by the dialectic between estrangement and cognition. The first term refers to the creation of an alternative fictional world that, by refusing to take our mundane environment for granted, implicitly or explicitly performs an estranging critical interrogation of the latter. But the critical character of the interrogation is guaranteed by the operation of cognition, which enables the science-fictional text to account rationally for its imagined world and for the connections as well as the disconnections of the latter to our own empirical world. If the dialectic is flattened out to mere cognition, then the result is “realistic” or mundane fiction, which can cognitively account for its imaginings but performs no estrangement; if the dialectic is flattened out to mere estrangement (or, it might be argued, pseudo-estrangement), then the result is fantasy, which estranges, or appears to estrange, but in an irrationalist, theoretically illegitimate way.

      This definition seems to me not only fundamentally sound but indispensable. Yet in Suvin’s own formulations the concept of science fiction as the fiction of cognitive estrangement involves at least two serious problems—both of which, however, may well be mere inadvertencies and both of which can in any case be solved within the basic Suvinian problematic (which can itself thus be enriched).

      The first problem is that the category of cognition appears to commit the literary critic to making generic distinctions on the basis of matters far removed from literature and genre. The awkwardness does not transpire so long as we are thinking, say, of Heinlein’s The Man Who Sold the Moon (1950) as paradigmatic of (cognitive) science fiction and Tolkien’s The Lord of the Rings (1954–1955) as paradigmatic of (noncognitive) fantasy. The rational connections that link D. D. Harriman’s world to our own are clear and direct, while the evident absence of any such connections between our world and that of the hobbits and orcs is equally clear. Yet there is a great deal of literature—some of it commonly labeled science fiction, some commonly labeled fantasy, and some, significantly, labeled both—that is based neither on the careful, straightforward extrapolation of Heinlein’s novella nor on the sharp break with known empirical reality of Tolkien’s trilogy. Joanna Russ’s The Female Man (1975) is considered science fiction, but few physicists would unhesitatingly affirm that the notion of parallel universes on which Russ’s novel depends is a valid cognitive option. Must we wait for a scientific consensus on the matter before deciding whether the text is science fiction or fantasy? H. P. Lovecraft has been described both as a science-fiction writer and as a writer of horror fantasy. Do “The Shadow over Innsmouth” (1936) and “The Dunwich Horror” (1939) earn the title of science fiction because their monstrosities have their origin not in the admitted supernatural but in vulgar pseudo-Darwinian notions of racial degeneration? What of C. S. Lewis’s Out of the Silent Planet (1938) and the following two novels in the Ransom trilogy? If theology is a science (if, to put it bluntly, Christianity is true) then the powerful estrangements produced by Ransom’s adventures on Mars are wholly cognitive; if religious dogma, however, is in fact as precritical as most critical theorists would insist, then Lewis’s epistemology is not really cognitive at all.

      All these examples suggest that cognition proper is not, in the strictest terms, exactly the quality that defines science fiction. What is rather at stake is what we might term (following a familiar Barthesian precedent) the cognition effect. The crucial issue for generic discrimination is not any epistemological judgment external to the text itself on the rationality or irrationality of the latter’s imaginings, but rather (as some of Suvin’s language does, in fact, imply, but never makes entirely clear) the attitude of the text itself to the kind of estrangements being performed. Comparison between Lewis and Tolkien is especially illuminating in this context, because both trilogies are concerned with conveying almost precisely similar orthodox Christian values. The Lord of the Rings is understood as fantasy and Out of the Silent Planet and its sequels as science fiction: not because it would necessarily be less rational to believe in hobbits and orcs than in planetary angels and Merlin redivivus, but because of the formal stances adopted by the texts themselves. Tolkien’s trilogy proclaims in its very letter a noncognitive disjunction from the mundane world (the kind of disjunction in fact suggested by Tolkien’s own central critical category of literary production as “sub-creation”),²⁰ while Lewis’s trilogy considers that principles it regards as cognitively valid cannot exclude events like the action fictionally portrayed from occurring within the author’s actual environment. Lewis, accordingly, produces a cognition effect, while Tolkien quite deliberately does not.

      Unless the distinction between cognition and cognition effect is kept steadily in view, the definition of science fiction as cognitive estrangement can lead to patent absurdities. For example, one of Asimov’s science-fiction mystery stories (“The Dying Night,” originally published in 1956) depends for its plot resolution on the assumption that Mercury has a “captured” rotation; that is, that it turns on its axis at precisely the same rate that it revolves around the sun, and therefore that it contains areas where night is permanent. This assumption was faithful