Recognizing science fiction

2.1 What is in a name?

Of all the generic narratives, SF is the most difficult to define. Romance, thriller, and zombie narratives, for example, are not only instantly recognizable but also readily defined by their necessary conditions. SF, however, need not be necessarily scientific. The problem is that everyone recognizes SF but no one knows what it is. As remedy, critical projects on SF often begin with an attempt at or rejection of a definition coextensive with an acknowledgment of the difficulty thereof:

The obvious place to begin a Critical History of science fiction is with a definition of its topic, but this is no easy matter. Many critics have offered definitions of SF, and the resulting critical discourse is a divergent and contested field. (Roberts 2016: 1)

What is science fiction? A weird, popular genre full of spaceships, laser guns, robots and bug-eyed monsters? Fiction concerned with the impact of science and technology on human and social life, and thus the literature best suited to understanding the contemporary world? A marketing category to be avoided if a text is to be treated as ‘real’ literature? The answer is in fact far more complex. Indeed, this book is premised on the idea that there is no such thing as SF, but instead multiple and constantly shifting ways of producing, marketing, distributing, consuming and understanding texts as SF. (Bould and Vint 2011: 1)

Attempting to define SF is, in this light, complicated. It would, nevertheless, seem important that projects, like this, build an apparatus of inclusion and exclusion whereby specific texts intrinsically become its core target and others clearly not. To this end, attempts include:

SF is, then, 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. (Suvin 1979: 7–8)

SF is that species of storytelling native to a culture undergoing the epistemic changes implicated in the rise and supersession of technical-industrial modes of production, distribution, consumption and disposal. It is marked by (i) metaphoric strategies and metonymic tactics, (ii) the foregrounding of icons and interpretive schemata from a collectively constituted generic ‘mega-text’ and the concomitant de-emphasis of ‘fine writing’ and characterization, and (iii) certain priorities more often found in scientific and postmodern texts than in literary models: specifically, attention to the object in preference to the subject. (Broderick 1995: 155; Roberts 2016: 1–2)

A charming romance intermingled with scientific fact and prophetic vision. (Stableford, Clute, and Nicholls 1993: 311–314; quoting Gernsback; Roberts 2016: 2)

Science fiction is anything published as science fiction. (Stableford, Clute, and Nicholls 1993: 311–314; quoting Spinrad; Roberts 2016: 2)

Definitions abound and suffer from falsification. Verification, here too, tends to invite complications. For example, Suvin’s classic definition of SF as “the literature of cognitive estrangement” problematically fails to verify the most widely recognized SF, namely Star Trek and Star Wars, as SF, whereas the former relies on telepathy and godlike beings and the latter on a supernatural “force” and a ghostly afterlife, features which suggest the fantastic (Suvin 1979: 4; Freedman 2000: 19). Other famous examples of quintessential SF equally tend away from the rational. For instance, in Solaris (1961), the entity covering the planet’s surface is an almost godlike consciousness that creates gargantuan babies, physics-breaking manifestations, and memory ghosts. Better still, the popular Dune series (1965–) has genetic memory tracing into the matrilineal past that can be shared by touching foreheads among the women who can survive an overdose of the poisoned water expelled by a drowned desert worm that, when alive, excretes a waste substance that people harvest, process, then ingest for longer lives, even though considerably longer life is attainable by the God Emperor, Leto II, through wearing the larval vector of these sandworms until he can see into the future but not his death because he will eventually breed people to not be seen by oracles, among whom number powerfully-enough his now-dead father who lives in his distributed worm-brain as an ancestral memory ghost. The fantastic, unnatural, and impossible too often creep too closely.^[4] What happens when Heinlein’s Stranger in a Strange Land (1961) or Le Guin’s The Left Hand of Darkness (1969) are not SF but Jurassic Park (1990–) is? Using theory either for verification or falsification against SF products qua ends becomes too problematic when it dispenses indispensable narratives. Nevertheless, if theory can apply to process qua means and attach to intent, then retaining Suvinean theory is as justifiable as it is advantageous. It merely requires a shift in perspective.

2.2 Mode

SF could be considered a species of the fantastic, or even necessarily fantastic (Parrinder 2001: 38), thereby short-circuiting the need to discuss fantastic complications. As analogies (Suvin 1979: 76) or descriptive metaphors (Le Guin 1979: 156–159), creative license could apply. Purists could even reject narratives with any nonzero percentage fantasy, thereby eliminating “nine-tenths of the scientific romance and pulp fantasies that most readers consider prototypes of the genre” (Csicsery-Ronay 2011: 139; after Freedman 2000: 18). SF could be framed more inclusively as the parabolic synthesis of creation and signification through “Cawelti’s work on popular formulas, Samuel R. Delany’s descriptions of science fiction’s reading protocols, Gary K. Wolfe’s study of sf icons, Joanna Russ’s ideas about sf subjunctivity, Hamon’s theory of the megatext, and William Tenn’s description of sf writers as jazz musicians riffing on one another’s work” (Attebery and Hollinger 2013: ix). Or, it could be considered historically, as a function of “production, distribution, and reception” across a “shifting set of conventions and expectations” that “influence…an intersecting but heterogeneous array of practices” (Rieder 2017: 161). Possibilities abound. Issuing therefrom, particularly for critical approaches, either Suvin’s classic definition should remain, be revised, or be dismissed—and each potentiality comes with costs, camps, and chances, all astonishing.

I suggest that it remain, but we should frame it differently: not as a definition of things but as an understanding of the process that produces them, a matter of how not what. We should not treat Suvin’s conceptualization as a falsification device that evaluates narratological ends but as a verification device that detects attempted means: a novologically estranging process during which the narrative becomes science fictional. Just as evaluating justice based on racist judges reveals injustice to be an end fault but not an intrinsic means fault, so too does evaluating SF by narratological products as SF differ from evaluating the intentional process of SF creation. This allows the eventual narratological end of SF to deviate, by authorial hand, without altering the essential means by which it became SF: the cognitively estranging novum.^[5]

If SF is a creative mode, then Suvin’s theoretical framework of cognitive estrangement through a novum can be classed as the process by which a narrative becomes generic SF, for “fictions and fabulations…are both processes of speculative extrapolation,” which maps to means, and, as follows, “testing to see whether they work or not, and what consequences follow from them” map to ends (Shaviro 2016: 10). This understanding connects reasonably well with Le Guin’s stance that one might read “a lot of” SF “as a thought-experiment,” for if “the purpose of a thought experiment…is not to predict the future…but to describe reality, the present world" and if "science fiction is not predictive; it is descriptive” (Le Guin 1979: 156), then creative thought-experiments are modal functions of cognitive estrangement, in that the novalogical situations of thought-experiments are estranged from reality and require the cognitive (re-)consideration of possibilities.^[6] In this way, “speculation,” “extrapolation,” “fabulation,” and “thought-experiments” attach to novalogical creation and “consequences,” from descriptions that inspire reflection to deviations into the fantastic, unfold afterwards. Just as an automobile factory fashions vehicles into types thereof according to a process, if the author qua driver forces the vehicle into a pond, then it is not a fault of the process engineering or the specifications thereof. The original intent, here, matters more than the result. Customization can void the warranty. Whereas all narratives begin at nothing and become something, the creative process to establish novalogical structuring structures for cognitive estrangement is simply “a way of getting something done,” which is definitionally a mode (Hollinger 2014: 139–140). Mode offers a way out of verifying texts against a monolithic model by replacing it with a method: not science fiction, per se, but a “science fictionality,” after Csicsery-Ronay (2011), who argues it to be “a way of thinking about the world, made concrete in many different media and styles” to the extent that it is “essentially a style of estrangement and dislocation” that produces a “kind of awareness we might call science-fictionality, a mode of response that frames and tests experiences” (ix–2), which, through the dialectic of cognition and estrangement “[allow] us to not only recognize the world of the story but also to see it as strange, prompting creative understanding and critical reflection about the difference between the text’s world and our own” (Vint 2014: 38). Thus, we develop a theory of science-fictionality: modes of creation and interpretation dancing dialectically.

Considered this way, science fictionality and creation thereof qua mode not only frees SF from structural attempts to entirely define it extrinsically (as something) but also commits it to the realm of individual responses intrinsically within social production (as the work of something [i.e., a process]). When experiencing a generic narrative recognized as SF, therefore, we are not noticing the current presence of modus operandi but the earlier science fictional moves thereof (qua Salthean development) previously executed within the generative phase of a narratological enterprise. In other words, we recognize the novalogical elements qua prototypes that commit the narrative to the SF genre faster than the possible fantastically derailing deviancies thanks to our culturally-tuned theory of SF. Simply, the estranging elements execute signifying structures faster than the plot. To the extent that SF narratives are not only “different in time and place from our own” but also primarily explore “precisely the difference that such difference makes” (Freedman 2000: xvi; quoted in Vint 2014: 49), this must be the case (i.e., the process of creation foreruns the process of exploration). Thus, when Suvin suggests that “SF is, then, 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” (1979: 7–8), nova become the modal arenas in which cognition and estrangement both present and interact. In this way, when how signals differently than what the difference of the difference attaches to what and the “difference which makes a difference,” that is, information (Bateson 2000: 315), attaches to how. As follows, when categorizing texts ranging across time, say, from the golden age to the new wave to now, comparisons become possible because of differences that attach to “what” (e.g., voyage, militaristic, identity, economies, genders, race, etc.). However, viewing from “how” collapses differences into prototypes (e.g., bifurcations of setting, timing, technology, population, etc.) of alternative frameworks.^[7]

The ontological commitment to arena over operations therein, means over ends thereto, possibility over authorial actualization thereof runs so counterintuitively against how we would traditionally verify generic texts that there must be a reasonably persuasive reason therewith: the recognitional elements (namely, prototypes, exemplars, and theories) emerge only from a creative modality of estrangement through development as functions of predictability and irreversibility. If prototypes and theories allow us to recognize SF as SF, then we are relying on the modalities that give rise thereto and the cognitive frames (Berlatsky 2009; Lakoff 1987, 2014) we build thereon. Since SF is recognized, considering, we have an opportunity to imagine new possibilities for conceptualizing estranging aspects of the genre through neuroscience more than a singular understanding of cognition, “a process that interprets information within contexts that connect it with meaning” (Hayles 2017: 21).

3.1 Cognitive architecture of recognition

Recalling the previous section, the moment in which recognition occurs must rely on an operation of (1) and (2) in which (2) returns a truth value. Such an operation suggests that there must be a framework controlling the operation. In other words, recognition is a superior cognitive function with two subordinate operations. Accordingly, I proceed from the notion that the operation of recognition is a supervening function of cognition. This commitment requires recognition to establish a cognitive framework to structure, sustain, and carry out its operations. Toward that end, recognition can be seen, functionally, through a cognitive architecture, where such semiotically model mental states as systems and “theories of cognitive architecture” explore “the nature of the basic structures and processes involved in cognition” (Aizawa 2007: 172). For my purposes, I contend that Durst-Andersen’s semiotic model best represents mental states of recognition, in that: “first you have a perceptual input, then you get a cognitive intake by involving background knowledge and finally you end up by having a perceptual-cognitive outcome,” which maps to three states: input (“the stage of experience”), intake (“the stage of understanding”), and output (“the stage of storing”) (Durst-Andersen 2011: 66). In this way, recognition is an inferential mental state—an atomic thought proposition (outcome) that is precipitated by a sensorial, perceptual experience (input) that is cognitively reasoned to be identical to, similar to, or typical of a previous perception (intake), as seen in Figure 1.

Figure 1:

Recognition model.

This occasions two orders of organization—(1) a superior framework buttressing and eventually combining (2a) and (2b) into a new mental state, (2c [recognition]), in this case, reason; (2) a subordinate series of mental states where (a) the first processes the present perception, (b) the second executes a recall function exploring the memories of past of perceptions, and (c) if the recall function is successful, then recognition, else it is unrecognized.

3.2 Modeling concepts

Across cognitive science, scholars have developed various theoretical ways to structurally model concepts, including the following, which attach to SF in the following ways explored in Table 1.

Each system has its advantages and disadvantages, but I submit that the dual theory carries the greatest potential for recognizing SF because it can rely on both (prototypical) fast impressions and (theoretical) reasoned considerations supported by real-world relations. After a fashion, while certain features of a hotel key direct you to which hotel the key belongs (qua prototypes), only the right notches and ridges against the right pins will open the right door (qua theory). This approach allows the public and scholars to recognize (or reject SF) either culturally or formally. Ultimately, though, it democratizes and personalizes one’s knowledge of SF, whereafter people can recognize SF, whether it deviates from academically classical definitions, as ends, while relegating the productive means that created it to verification and criticism.

3.3 Modeling SF as a recognizable concept

Section 3 first builds a cognitive architecture for recognition then builds a cognitive model of concept recognition. Both systems must now functionally be joined for the recognition of SF as a concept to work. Whereas recognition is an inferential mental state precipitated by a sensorial, perceptual experience that is cognitively reasoned to be identical to, similar to, or typical of a previous perception, to the extent that recognition of concepts is possible, and to the extent that both models can be represented structurally, they could be integrated. This subsection merges the cognitive architecture of recognition (Section 3.1) with the two cognitive models of concepts that work with SF (Section 3.2) to structure recognition of SF.

Just as the cognitive architecture of recognition scans the past for an encounter with the experience engaged in the present, so too do we see the reliance on the past for cognitively modeling concepts. There are enough common features for a shared structure to recognize concepts between models of recognizing and concepts such that I can map both onto the same cognitive architecture for recognizing SF in Figure 2.

Figure 2:

Recognition of SF through dual theory of prototypes and theory–theory as supported by exemplars.

One cognitively models then recognizes SF qua concept through experience. First, one must encounter SF that is sustained through a mechanism to learn it as a narratological category, whereupon one stores it in memory as an exemplar of SF. Second, one must again experience SF, whereupon its probability, typicality, and tendency to the previous perception contribute to and are quickly assessed against the emerging prototypes built on the exemplars experienced, while, simultaneously, the same exemplars contribute to and are measured by a newly coalescing theory of SF qua frame (Bowman and Zeithamova 2018: 2612; Grishakova 2009: 189; Medin and Schaffer 1978: 209; Murphy 2002: 74; Nosofsky 1988: 707; Rosch and Mervis 1975: 574). This process, of course, follows the Salthean model of development and evolution, such that predictability attaches to prototypicality, irreversibility to learning, and the accumulation of historical information (qua Batesonean difference) to the theoretical. The model will continue to refine the prototypes and the theory based on experience, whatever the modality (Bauer and Just 2019: 531; Musz and Thompson-Schill 2019: 565), naturally through clustering, for: (1) the more frequent the encounter with a category, the more we learn “about its members, including about their perceptual properties” (Weiskopf 2015: 241), which aligns with our understanding of the interplay between the fast perception of prototypes, nonconscious cognition, and cognition to the extent that “the processes nonconscious cognition uses to discern patterns are constantly in motion, reaching metastable states as patterns are discerned and further reinforced when temporal matching with the reverberations between neural circuits cause them to be fed forward to consciousness” (Hayles 2017: 24); and, (2) tying into Peircean semiosis and the theory–theory, “Sf images become iconic through repetition” (Attebery 2013: 9). SF is, thus, learned through recognition. To reinforce the foregoing ways in which SF is learned to be recognized the following section addresses how learning models align with concept theories.

3.4 Learning categories through models

Across cognitive science, scholars have developed various theoretical ways to model how concepts are learned structurally and through which areas of the brain, including the following, which map to the foregoing theories and attach to SF in the following ways explored in Table 2.

Just as the recognitional model requires a dual theory for prototypes and theory, so too does learning the concept of SF require the same, which only suggests the clustering model. It is, therefore, most likely that we learn SF in the same way we learn trees without definitions: prototypes through exemplars. To the extent that prototypes and exemplars activate multiple systems, then for our purposes, the clustering model should be seen, ipso facto, as a multiple systems model. Thus, when we say that SF is cognitive, that it is through cognition that we view SF, we must remember that multiple systems are activated by learning SF and that, beyond cognition, we also have recognition. When we socio-cognitively learn SF by its exemplars and the prototypes that emerge from estranging nova and the theories we build thereon, we learn SF through multiple pathways that connect through and to the hippocampus. The related systems track and process visualizations, contexts, rules, exemplars, prototypes, and clusters. We need not exclusively look to SF’s narratological ends to discuss cognition. Indeed, we can, instead, discuss novalogical excitement across multiple systems of consciousness (after Hayles 2017). Thus, while we learn SF through prototypes, which are supplemented by exemplars that cluster to permit a theory to develop that we can summarize into features—howsoever humanly imprecise, imperfect, and individualistic—it is not only the domain of cognition any more than trees could ecologically be the forest.