Digital twins as space media

Abstract

NASA scientists have claimed that the rescue of the stricken Apollo 13 spacecraft in 1970 created the first digital twin system. In this article, I examine NASA’s narrative, which is gathering consensus as the origin story of digital twin technologies, with two main points of focus. First, through an examination of the dynamics of control and error, I scrutinise the ways in which the NASA ethos and worldview may be embedded within contemporary uses of digital twin technologies. Second, through an analysis of two subsequent NASA missions and the 1968 film 2001: A Space Odyssey, I frame digital twin technologies as ‘space media’ which, Gál et al. claim, can ‘reorient our view [. . .] to “off-Earth” perspectives’. This approach reveals the potential of digital twins to propagate imaginaries beyond the managerial systems of control that constitute their dominant contemporary usages.

Introduction

On 14 April 1970, 56 hours after its launch from the Kennedy Space Centre and 200,000 miles from Earth, damaged wiring created a spark inside one of the liquid oxygen tanks aboard the Apollo 13 spacecraft. Fuelled by the oxygen, fire spread and rapidly increased pressure, causing an explosion that blew out a 13-foot panel of the fuel cell bay, exposing the innards of the service module to the abyss of space. What happened next is well known, immortalised by the recordings of dialogue between the crew and mission control, the crew’s testimonials delivered in carefully choreographed press conferences, and, for later generations (including my own), Ron Howard’s 1995 Hollywood blockbuster, Apollo 13. By the time the spread of damage had been stemmed, the service module had lost two of the three fuel cells that it had been relying on to take it to the Moon and back. With its original plan of landing by the Fra Mauro crater abandoned, the NASA team refocused efforts on safely returning the three crew members to Earth. To do so, engineers back in mission control combined digital modelling using live data streams from the in-flight aircraft with physical training simulators to test possible solutions to problems as they emerged. While simulating possible conditions of spaceflight had long been central to NASA’s mission preparation, it is the live feedback loop between simulation and spacecraft that makes Apollo 13 the first digital twin.

This is the narrative that has been gathering consensus as the origin story of digital twin technology. In a keynote speech at the 2021 Digital Twin Summit, Allen, a Senior Technologist for Intelligent Flight Systems at NASA, argued that the idea of a ‘digital twin’ was born during the Apollo mission of the 1960s, through the creation of ‘living models’ of the spacecraft, and first properly put into action in response to the Apollo 13 failure. As Allen puts it,

NASA employed multiple simulators to evaluate the failure and extended a physical model of the vehicle to include digital components. This ‘digital twin’ was the first of its kind, allowing for a continuous ingestion of data to model the events leading up to the accident for forensic analysis and exploration of next steps.

In their ‘comprehensive review’ of digital twin scholarship and applications, Thelen et al. (2022: 4) also turn to NASA for the first definitions of the concept, citing a 2012 conference paper in which NASA and U.S. Air Force scientists Glaessgen and Stargel (2012: 7) gave the first specific definition of digital twins in the aerospace domain as: ‘an integrated multiphysics, multiscale, probabilistic simulation of an as-built vehicle or system that uses the best available physical models, sensor updates, fleet history, etc., to mirror the life of its corresponding flying twin’. Singh et al. (2021: 3) build on NASA’s definition and examples to propose their description of digital twins as being ‘a virtual representation of a physical object’ in which there is a ‘bidirectional transfer or sharing of data between the physical counterpart and the digital one, including quantitative and qualitative data, historical data, environmental data, and most importantly, real-time data’.

Most scholars of digital twins, writing from an engineering perspective (including both Thelen et al. and Singh et al.), briefly mention the background in aerospace technology both as evidence of the concept’s distinctiveness from related systems of simulation and modelling and as proof that contemporary iterations are building on legitimate precursors that were developed within respected engineering cultures. By contrast, in this article, I examine the digital twin origin story with two main aims: first, to scrutinise the ways in which the NASA ethos and worldview may be embedded within contemporary uses of digital twin technologies; and, second, to propose how thinking of digital twins as ‘space media’ can help us to think differently about a technological approach that is becoming so central, through its application to managerial and production systems, to the digital infrastructures of our daily lives. In their survey of critical engagements with the concept of outer space at the intersections of Science and Technology Studies and Media Studies, Gál et al. (2021) use the term ‘space media’ to address how space travel and exploration have had ‘a formative role on some of our media configurations’ (p. 654), from the development of satellite television as a mass medium to the concept of the cyborg. At the same time, they argue, our understandings of outer space – as well as ‘the limits of the outer, the outside, the external’ more generally – are ‘defined by the limits and constraints of media’ (Gál et al., 2021: 658). By simultaneously extending humanity’s epistemological mastery and pointing to the vast universe that exists beyond its limits, space media are characterised by a fundamental ambiguity. For instance, Kessler (2012) has argued that, by invoking aesthetic traditions of the Romantic sublime, images produced by teams of scientists working with the Hubble telescope ‘encourage the viewer to experience the cosmos visually and rationally, to see the universe as simultaneously beyond humanity’s grasp and within the reach of our systems of knowledge’ (p. 5). Due to their centrality to dominant understandings of the cosmos and therefore our fundamental epistemological assumptions, thinking through space media has the capacity to ‘reorient our view’ to how ‘off-Earth’ perspectives serve to both reinforce and challenge our ‘on-Earth’ (Gál et al., 2021: 657) beliefs and the socio-technical systems that support them.

While engaging closely with the field of space media studies outlined by Gál et al., this article takes inspiration from Jue’s (2020) study Wild Blue Media, in which she uses a ‘science fiction strategy’ (p. 5) to frame the ocean as a medium that functions as a ‘disorientation device [that] denatures our normative habits of orienting to the (terrestrial) world’ (p. 6). Similarly, I propose that thinking of digital twins as space media implies taking seriously the ‘bidirectional flow of data’ that is constitutive of the technology and engaging not just with how the Earth-bound twin binds the space-bound twin to terrestrial norms but also how the celestial twin pulls its earthly sibling out of its epistemological conventions and assumptions. After examining the dynamics of control and error that constituted the conditions of emergence for the ‘first’ digital twin in the Apollo 13 mission, I discuss three examples of digital twin dynamics playing out in the context of space media: two taken from more recent NASA missions, the Mars Exploration Mission and a Twin Study conducted in 2015–2016; and one taken from Stanley Kubrick’s highly influential science fiction film of 1968, 2001: A Space Odyssey. In each case, twins display the ambiguity that is fundamental to space media; their objectives of prediction and control are set in tension with counter-tendencies that open onto alternative epistemological and ontological perspectives.

Cybernetic control and error

If we take the NASA narrative seriously, then the first digital twin emerged in response to failure. The Apollo 13 mission has become inextricable from the aesthetics of failure within mass-cultural Anglophone imaginaries. Command module pilot Jack Swigert’s supremely understated communication with the NASA ground control – ‘Houston, we’ve had a problem’, which was changed by Apollo 13 scriptwriter William Broyles to ‘Houston, we have a problem’ – has become a commonplace phrase for announcing the sudden appearance of seemingly insurmountable challenges. Despite the clear lineage within NASA, Allen (2021) is also echoing the logic of contemporary tech companies who reframe system failure as an error that is a stepping stone to progress. As Thylstrup (2021: 192) puts it,

within the experimental epistemology of tech companies, each new technology that appears on the back of a failed experiment becomes yet another stage in an ever-unfolding tech metamorphosis in which each new demo sheds its old skin to unveil a newer, larger, and more successful one.

Regardless of NASA’s rhetorical aims, approaching digital twins in the context of failure is revealing. Drawing on a long tradition of viewing breakdown and failure as ‘the means by which societies learn and learn to reproduce’ (Graham and Thrift, 2007: 5) – much of which can be traced back to Star (1999) – Parikka (2016: xxxiii) argues that digital technologies that have been naturalised to the point of ‘ontological invisibility’ only reveal themselves in the event of breaking down. Accidents, he argues, ‘reveal technology, and the power/knowledge relations that media are embedded in’ (Parikka, 2016: xxxiii). As a response to the failure of Apollo 13, digital twin technology embodied fundamental aspects of NASA’s approach. Through an analysis of the structures and procedures in place during the Apollo mission, Sage (2014: 11) argues that NASA functioned as an ‘idealized technocrat organization’ that combined ‘managerialist discourses of efficiency and rationality’ with ‘transcendental myths of exceptionalism and destiny’. The model of ‘rationalistic’ and ‘elitist’ organisation that Apollo helped ‘incubate’, which came to be known by the term ‘technocracy’, ‘assumes that new technology, scientific and engineering expertise, big business and big government, can gather together around logics of efficiency and cost-effectiveness to solve an array of economic, environmental, social, and even moral problems’ (Sage, 2014: 57).

As Sage (2014: 61) points out, within this model, there was no room for ‘ambiguity, for messiness, for wasted time, for unknown space’, all of which were to be ‘translated into calculated risks’. Like similar malfunctions before and since (such as those that led to the Apollo 1 and Challenger disasters of 1967 and 1986), the Apollo 13 malfunction represented not only a failure of the mission but also a challenge to the entire NASA approach. The safe return of the three crew members to Earth has enabled subsequent mythologising to claim the mission as the ultimate triumph of NASA engineering and management. In a moment that encapsulates the ethos of the 1995 film, for instance, when a senior member of mission control articulates a fear that it might be ‘the worst disaster NASA has ever experienced’, Chief Flight Director Gene Krantz (played by Ed Harris) interrupts him to say: ‘with all due respect, Sir, I believe this is going to be our finest hour’. The fact that the NASA engineers found technical solutions to the technical issues experienced by the spacecraft reinforced the ‘utopian attitude towards technology’ that, as Sage (2014: 57) points out, lies at the heart of NASA’s technocratic model. The digital twins approach represented an intensification of the control that NASA’s Earth-bound engineers had over the in-flight spacecraft. Continually updated with live data, the digital model could function as a prediction machine, enabling technicians to foresee and manage future developments. The ever-growing number of sensors and data points transmitting from the spacecraft turned capsules into spaces of almost perfect surveillance. Damjanov and Crouch (2020: 10) point out that within the International Space Station (ISS), which was launched in 1998, ‘all activities [. . .] are continually monitored through multiple forms of scrutiny, which include comprehensive details about the status and operations of all its systems and subsystems’ fed by what NASA claims to be ‘approximately 350,000 sensors’. Seemingly separate, the ISS is in fact tethered to its ground control twins with which it is ‘closely interlinked’ through constant ‘signal traffic’ (Damjanov and Crouch, 2020: 7).

The Apollo 13 failure also brought to light the constitutive interconnections between human and non-human elements in space capsules. Given the depletion of fuel on board the service module, the digital twin was initially used to model possible strategies for using the lunar module to complete the voyage home, despite the fact that it was designed only to take two crew members from lunar orbit to the surface of the Moon and back. Attention then turned to how to adapt the life support system on board a lunar module that was not designed for three crew members. This included managing the use of what NASA terms the ‘consumables’ that are essential for maintaining human life: oxygen, water, and electrical power. Achieving this involved a combination of digital modelling (for instance, testing landing sequences that would use the least amount of power possible) and more analogue experimentation, such as repurposing parts of the Extravehicular Activity (EVA) suits and spare gaffer tape to patch the service module’s lithium hydroxide cannisters into the lunar module’s life support system. The latter was necessary to ‘scrub’ enough carbon dioxide from the air inside the lunar module to maintain its breathability for the duration of the return journey. The requirements for returning the aircraft to Earth had to be carefully balanced against maintaining the necessary conditions for human survival. The first digital twin was modelling not just the engines and electrical equipment but the entire ecosystem of the capsule, including the many points of interface between human crew and their technological environment.

Digital twinning systems emerged, therefore, as a key component of what Anker terms the ‘ecological colonization of space’. ‘Space ecology’ is the term Anker (2005: 241) uses to refer to the intersections between ecological thinking and the ‘colonialist agenda of space research’. During the 1950s and 1960s, a number of researchers used ecological models for thinking about how to sustain human life within space capsules. Best known among these were Howard T. Odum’s designs for a space cabin as a ‘self-regulating cybernetic system with the ability to maintain the chemical components of the atmosphere through negative and positive ecological feedback loops that provided comfortable living conditions for the astronauts’ (Anker, 2005: 246). In a paper published in 1960, Clynes and Kline (1960: 26) used the concept of the cyborg to think about how cybernetic research can help ‘the task of adapting man’s body to any environment’, including the inhospitable environment of space. ‘The Cyborg deliberately incorporates exogenous components extending the self-regulatory control function of the organism in order to adapt it to new environments’ (Clynes and Kline, 1960: 27). Although Odum’s proposals were not directly taken up, Anker demonstrates how his focus on human inhabitants as part of the technological system was influential on technocratic social management. ‘His methodological reductionism of all biological life (including human behavior) to charts of energy circuits became the justification for proposals to manage human society scientifically’ (Anker, 2005: 246). Contemporary implementations of digital twin technology for the management of social systems, including smart cities, echo NASA’s dream of controlling the contingencies of human life through cybernetic prediction and management.

However, the role of failure in this narrative of the emergence of digital twin technology has a further significance. By drawing attention to the contingencies of technical systems, their material structures, as well as their assumptions and oversights, failures also point to how things might have been different. For instance, the designers of the lunar module never contemplated the possibility that it could be used as a ‘lifeboat’ should the main service module fail. The designers had also never contemplated whether the electronic systems in the lunar module would continue to work if the temperature dropped (due to conserving energy for re-entry) or when wet from condensation (due to low oxygen levels in the capsule). But, beyond the details of design decisions and their blind spots, failures can also draw attention to more fundamental assumptions behind technological systems and, in the words of Feenberg (2002: 19), the ‘understanding of human life [that] is embodied in the prevailing technical arrangements’. In the process, failures can point to different technological systems and logics that produce different ‘understandings of human life’. Addressing the context of ‘big data’, Thylstrup draws attention to the fundamental ambiguity of error. On the one hand, the identification of error, especially in normative conceptions of human identities, has always been fundamentally political and a central strategy of disciplinary power. On the other hand, the political status of error has, as Thylstrup (2021: 196) puts it, ‘given rise to a reengagement with error as an inroad for critical and subversive engagement with power’. From this perspective, errors ‘hold alternative potential, which can be used against hegemony by showing that alternatives are embedded in dominant cultures’ (Thylstrup, 2021: 196). One recent example of this approach to error is Legacy Russell’s work on what she calls ‘glitch feminism’. She argues that, by suspending intended functionalities and interfaces, glitches ‘create a fissure within which new possibilities of being become manifest’ (Russell, 2020). As a product of failure, the first digital twin provides a blueprint for the dominant uses of the technology in production and systems management while at the same time pointing to possible alternatives. In this, it embodies the epistemological ambiguity characteristic of space media.

Another way of approaching this is to examine how digital twin technologies both echo and depart from the logic of proxies in epistemological systems. Mulvin (2021: 20) has argued that ‘proxies’, understood as the standards shared by communities of knowledge (such as the International Prototype Kilogram), mediate between an unpredictable and chaotic world ‘out there’ and the ‘controlled space of knowledge production “in here”’. As such, they simultaneously shore up the boundary between the two and point to the fact that it is ‘always a permeable fiction’ (Mulvin, 2021: 20). Digital twins differ from the epistemological logic of the proxy in many ways. For example, unlike the illusion of universality and consistency over time provided by the proxy, digital twins construct a standard that is both individualised and cybernetically updated to reflect changing realities (albeit according to a logic of ‘optimisation’ that is assumed as a universally desirable goal). However, they do echo proxies in their function as intermediaries in the sense outlined by Mulvin. Their first usage, as we have seen, was to mediate between NASA’s technocratic system of governance and the inhospitable and largely unknowable vastness of space that continually threatened to engulf it. While the intended usage of the digital twin was to contain the latter by the former, we can also approach digital twins as a form of space media that opens up terrestrial knowledge systems to the mysteries of space and troubles dominant epistemological structures. Sage (2014: 72) points out that some of this ambivalence was always a part of the Apollo project since ‘NASA’s technocratic triumph [. . .] was always bound up with something beyond itself: transcendental spaces and times which are far from calculable in technical abstractions’. In the following two sections, I examine subsequent examples of the technology in the context of space travel and exploration that embody this counter-logic to the dominant uses of digital twin technologies.

Mars exploration rovers

This counter-logic of digital twins is visible in other uses of twins or the concept of twinning in later NASA missions and projects. In this section, I explore two examples: the characterisation of NASA’s ‘twin’ Mars Exploration Rovers that were active between 2003 and 2018, and astronauts Scott and Mark Kelly’s twin studies experiment of 2015–2016. Both examples use twins to signal uncanny shifts in human perception and epistemological blockages produced by space exploration. As part of the Mars Exploration Program initiated in 1993, aimed broadly at reaching a clearer understanding of the potential for life on Mars, NASA launched two robot rovers, MER-A and MER-B, which landed at separate locations in January 2004. The rovers were six-wheeled robots around the size of ‘golf carts’ that were powered by solar panels and equipped with antennas for communication to and from Earth and nine cameras used to identify rocks and soils that might reveal information about the existence of water on the planet. Both robots were designed and operated from the Jet Propulsion Laboratory in California. An article published by NASA to mark the 20th anniversary of the robot rover landings in 2004 celebrates the achievements of the ‘twin’ robots, which included the discovery by MER-A of ‘spherical pebbles of the mineral hematite that had formed in acidic water’ and, by MER-B, of ‘signs of ancient hot springs’ that might have created the conditions for microbial life. The language and logic of twinning is evident in the rover mission in two main ways: the robots are characterised as identical ‘twins’ of each other, while the continual feedback loops between the Jet Propulsion Laboratory and the robots embody a digital twin logic.

First, the two robots were characterised as ‘twins’ by the laboratory in press communications, including the anniversary article. This characterisation is part of a broader process in which the team anthropomorphised the rover robots. In her ethnographic study of the project, Vertesi describes how, despite the fact that they were built as exact replicas of each other, the laboratory team started to identify differences between the robots and gave them names that reflected their distinct personalities:

The robots even have social class: during my ethnography, Spirit [MER-A] was frequently described to me as a ‘blue collar’ laborer who had to work for every success she earned, while Opportunity [MER-B] was a ‘golden girl’ who found evidence of water on Mars immediately after landing. (Vertesi, 2012: 399)

Their characterisation as twins echoes another pair of robots, Elmer and Elsie, that were developed and described as twins by British neurophysiologist and cybernetician Gray Walter (1961 [1953]) in the late 1940s. In his book The Living Brain, Walter argues that, despite their simplicity, the robots displayed unprogrammed emergent behaviour when they interacted with one another. As with the rovers, the description of the robots as twins served to anthropomorphise Walter’s two robots, a rhetorical move that reinforced his aim (which he shared with other cyberneticists of the era) of undermining the ontological differences between human and machinic organisms.

Furthermore, just as monozygotic twins start life as exact copies of each other when a single zygote divides in two, the rover robots were constructed from the same hardware and software. However, just as ‘identical’ twins diverge in their development from the moment of separation through their different interactions with environmental factors (from blood supply in the womb to life experiences), the robots accrued differences through their different roles in the mission as well as the vagaries of chance. (Spirit almost immediately stumbled upon its important discovery, while Opportunity had to ‘wait’ several years.) This echoes the metaphorical use of twins in Norbert Wiener’s (1989 [1950]) book The Human Use of Human Being, which was published two years after the more technical Cybernetics: or Control and Communication in the Animal and the Machine. In a chapter where he sets out his vision of organic life in terms of informational pattern rather than essence or soul, Wiener (1989 [1950]: 102) draws a comparison between the developmental divergences between monozygotic twins and between two ‘computing machines’:

[J]ust as the future development of these two machines will continue parallel except for future changes in taping and experience, so too, there is not inconsistency in a living individual forking or divaricating into two individuals sharing the same past, but growing more and more different. (Wiener, 1989 [1950]: 102)

Wiener here echoes Walter’s use of twins as a metaphor to dismantle essentialised differences between humans and machines.

There is, however, an ambiguity in the metaphorical deployment of twins by both post-war cyberneticians as well as NASA’s Jet Propulsion Lab, which again reinforces their characterization as space media. On the one hand, the rhetorical move anthropomorphizes machines, framing them with the familiar terms of human relations. In a news segment broadcast by the BBC, Walter is shown in his suburban home with the twin robots scooting around on the carpet like domestic pets. On the other hand, twins are used to unsettle assumptions about human identity as being distinct and superior to the technological world. This usage draws on a long tradition within literary and visual cultures of associating twins with the uncanny. According to Sigmund Freud’s account, the feeling of the ‘Unheimliche’ (or unhomely) takes those who experience it back to a point in the development of the psyche in which separations between the human subject and the exterior world of objects have not yet properly formed. It is, for Freud (1986 [1919]: 366), ‘a regression to a time when the ego had not yet marked itself off sharply from the external world and from other people’. By appearing to lack differences between each other, twins can induce in spectators a fear of a breakdown of the boundaries separating them from the world around them. It is this quality that made twins (alongside doubles and doppelgängers of other kinds) a common trope among creators of the Gothic and its influences.

This ambiguity is also present in the logic of digital twinning evident between the rovers and the Jet Propulsion Laboratory. The robots were controlled by a team of researchers based in Pasadena, California, who updated digital models of the Martian terrain to reflect new data transmitted from the rovers. These changing digital models informed future instructions in a constant feedback loop between researchers and robots. Vertesi describes how the exchange went beyond the merely informational as the sensorial capabilities of the rovers started to inform how the researchers experienced their own bodies. The ‘elision between human and robotic experience’ (Vertesi, 2012: 399) took place at the level of language (comparing the robot cameras to eyes, for instance) and embodied interaction with the data (e.g. by physically emulating the robot’s movements to ascertain angles and distances between camera and objects). While the overall aim of the digital twin logic was to incorporate the data transmitted by the rovers into the team’s epistemological systems, Vertesi identifies a counter-tendency in this embodied identification with the rovers which emphasises their alignment with the ambiguities of space media. The researcher–rover relationship is ‘in opposition to anthropomorphism: it is a technomorphic move in which team members take on the robot’s body and experiences as part of their practice and narrative of their work’ (Vertesi, 2012: 400). The twinning between researcher and rover, which integrates digital modelling with embodied emulation, enacts a reconfiguration of human senses with technological systems. Messeri (2016: 5) argues that planetary scientists such as the Mars Exploration Rover team are ‘literally world-builders’ who are ‘invested in questions of what it is like to be on other worlds’. In the hands of these scientists, digital twins become tools for thinking across scales (from the human to the inter-planetary) and what Messeri (2016) terms ‘inhabiting’ (p. 19) other planets by replacing the ‘placelessness’ of space with ‘planetary places’ as part of a new cosmology in which ‘Earth becomes part of a vast interplanetary network’ (p. 16). While place-making practices such as this risk reinstalling existing ‘hierarchies and exclusions’ (Messeri, 2016: 17), as evidenced by the current generation of ‘NewSpace’ entrepreneurs, they can also create space to reflect on and dislodge these hierarchies and exclusions in existing configurations of place. Echoing Walter’s robot twins, the role of the Mars Rover twins, as space media, is to install these alternative perspectives at the heart of the epistemologically familiar.

NASA twin study

A similar ambivalence is present in a Twins Study experiment carried out by NASA scientists and astronaut brothers Scott and Mark Kelly. Starting in March 2015, Scott Kelly spent a full year in low-Earth orbit on board the ISS while his ‘identical’ twin brother, retired astronaut Mark, remained on Earth. Sensing the opportunity for extra research data (and positive press attention for NASA) Scott proposed carrying out a year-long experiment on the human body’s adaptation to long periods of time spent in space, using his brother as a control. As the published study puts it, ‘[t]o understand the health impact of long-duration spaceflight, one identical twin astronaut was monitored before, during, and after a 1-year mission onboard the International Space Station; his twin served as a genetically matched ground control’ (Garrett-Bakelman et al., 2019: 145). The aim of the project was to expand NASA’s existing research on the physiological demands and impacts of spaceflight, which it has been carrying out for decades, to ‘include molecular information’ and ‘give direction for counter-measures and additional studies to be done in the future’ (Garrett-Bakelman, 2019). ‘Genetic, immune system, and metabolic functions are of particular concern given exposure to space radiations, restricted diet, reduced physical work requirements, disrupted circadian rhythms, and weightlessness’ (Garrett-Bakelman et al., 2019: 145). The published study, led by Garrett-Bakelman, sets out a couple of important caveats, which reveal that the experiment was just as much about maintaining good public relations as forging new research pathways. First, while the experiment lasted 1 year, future trips (including ‘transit to and from Mars’) could be as long as 3 years (Garrett-Bakelman et al., 2019: 145). Second, with only one test subject, ‘it is impossible to attribute causality to spaceflight versus a coincidental event’ and so the study ‘should be considered as hypothesis generating and framework-defining and must be complemented in the future by studies of additional astronauts’ (Garrett-Bakelman et al., 2019: 146).

The experiment can be viewed as a variation on the digital twin approach. Rather than the Earth-bound twin Mark functioning as a digital twin of his brother Scott, it is the normative image of human health in space constructed from the twin study that functions as a digital twin for future NASA astronauts. So, data produced by the twin experiment contribute to a set of biological parameters that define future expectations about the impacts of long-term space travel on the bodies of astronauts. The biological norm built on this will then be updated by data from future astronauts in an ongoing feedback loop. As with Apollo 13, the digital twin is a tool of prediction and control: predicting impacts on astronaut bodies and controlling, not just those bodies, but also the interfaces between them and the environments of the space capsules. This is an example of what Olson (2010: 170) describes as ‘space biomedicine’, which is ‘a form of environmental medicine that seeks to optimize and manage technically enabled human ecologies where life and environment are dually problematized’. The form of subjecthood produced by space biomedicine is ‘hyper-medicalized’ (Olson, 2010: 174) and always studied in the context of its many interfaces with the technological and natural environment. The focus is on the wider ecologies in space capsules, which Olson (2010: 174) describes as ‘a collectivity of living and non-living things in vital if not biologically generative association’.

Olson (2010: 171) argues that, through this focus, space biomedicine anticipates an emerging ‘ecobiopolitical’ regime of power that takes as its focus ‘human life conceived in its entanglements’. What the parallels between the Apollo 13 mission and the NASA Twin Study reveal then is that digital twins have always been tools of ecobiopolitical control. As we have seen, the focus of the ‘first’ digital twin was the maintenance of the whole Apollo system, including the three crew members, just as the twins experiment focused on the entanglements of Scott Kelly with the environment of space and the life support system of the ISS. The research data contribute to what Olson (2010: 183) describes as an ‘evidence base of medical data on the environmentally contingent physiological changes that happen to spacefaring humans’, through which researchers construct a ‘space normal’, a model used to ascertain whether astronauts in space are ‘functioning within normal limits’. Through their role in ‘ecobiopolitical’ control, digital twins are therefore bound up with what Kilian (2021: 2017) describes as ‘the militarization of human nature’ during the Cold War, which entailed ‘bypass[ing] biological evolution by means of cybernetic control’ and ‘us[ing] military technology to adapt the human body to the hostility of the cosmos’.

However, as with the Mars Exploration Rovers, it is possible to identify a counter-tendency in the ecological focus of space biomedicine evidenced by the NASA Twin Study. Aronowsky (2017) proposes a different reading of NASA’s interest in systems ecology research, which Anker cited as evidence of ‘the ecological colonization of space’. A number of researchers followed Odum in turning to ecological thought for sustaining human life during long-duration space travel. These included proposals to add other species to the space capsule ecology. In 1958, for instance, Jack Myers proposed using algae-based life-support instruments that would recycle human waste, provide oxygen and constitute a source of nutrients. These studies (and NASA’s interest in them) present the space cabin ‘not as the expression of humans’ supremacy over their environment but instead as a space that foregrounds their vital reliance on it’ (Aronowsky, 2017: 361). From the embodiment of militarised masculinity captured in Tom Wolfe’s (2005 [1979]) The Right Stuff, the astronaut is re-signified as ‘enfeebled’ and ‘utterly dependent’; from an emblem of ‘technoscientific supremacy’, the space cabin is re-signified as a place of ‘interdependence’ (Aronowsky, 2017: 362). Astronaut subjecthood, from this perspective, ‘is a way of being in the world in which a sense of one’s self would be anchored in relationality, in which the act of inhabiting would demand a radical dependency’ (Aronowsky, 2017: 371). The Apollo 13 failure exposed the ‘radical dependency’ of the crew members on the life support systems of the service and lunar modules, momentarily puncturing the myth of individual heroism (even if that was subsequently reasserted by the NASA public relations machine and the Hollywood film). The digital twin model that emerged in response to that failure takes as its object the interrelations between living and non-living components. The NASA Twin Study, then, embodies the ambiguity of space media by being both a tool of ecobiopolitical control and an epistemological pathway beyond myths of individuality towards human dependency and relationality.

Science fiction imaginaries

Digital twins were conceptualised as space media in the world of fiction before the retrospective labelling of Apollo 13. Two years before the fuel cell failure on board the Apollo 13 service module, another error of space cabin infrastructure brought digital twins into the popular imaginaries of space travel. The third of the four sections of Stanley Kubrick’s 1968 science fiction epic 2001: A Space Odyssey, based on a screenplay co-written with Arthur C. Clarke, centres on the interactions between the crew of a spaceship called Discovery One, which is on a mission to Jupiter, and the onboard artificial intelligence HAL-9000. Unbeknownst to the crew, the aim of the mission was to investigate a possible intelligent presence connected to a mysterious alien monolith that had been found buried under the surface of the Moon. HAL, which is a shortened version of Heuristically programmed ALgorithmic computer, functions as ‘the brain and central nervous system of the ship’ and communicates with the crew in an eerily calm male human voice. Much of the exposition of HAL’s setup is carried out by a televised interview in which the computer claims to be ‘the most reliable [. . .] ever made’ since ‘no 9000 computer has ever made a mistake or distorted information’. However, as with Apollo 13 two years later, error is exactly what disrupts the smooth functioning of the mission when HAL’s Prediction Centre reports that the ship’s main communication link to Earth will be due to malfunction in 72 hours. When Dave Bowman and Frank Poole, the ship’s two conscious crew members (three others remain in induced hibernation for the duration of the trip), run diagnostics on the supposedly faulty unit, they find nothing wrong. Although HAL sticks to his original assertion (‘I don’t want to insist on it, Dave, but I am incapable of making an error’), Dave and Frank consult with Mission Control who claim that it is actually HAL’s Fault Prediction Centre that is in error. The reason they know this is that they have been running checks on HAL’s ‘twin’ back on Earth. ‘I know this sounds rather incredible, but this conclusion is based on results from our twin 9000 computer’. However, this narrative of events is never officially confirmed. Mission Control are in the process of ‘running cross-checking routines to determine the reliability of this conclusion’ when HAL renders its own prediction true by shutting down its connection with Earth and trying to kill the human crew, including those in hibernation.

The HAL plotline of 2001: A Space Odyssey, as well as the novel version that was published just after the film, is part of a tendency within Anglophone science fiction, which was already well established by the late 1960s, to use twins to explore the socio-technical implications of cybernetics research. Twins and doubles have long been employed in science fiction to emphasise the uncanny effects of technological and scientific developments on human subjectivity. In his 1956 novel, Time for the Stars, Robert Heinlein (2017 [1956]) constructs a narrative around the ‘Twin Paradox’, a thought experiment in special relativity which states that identical twins would age at different rates if one travels into space and back on a high-speed rocket while the other remains at home. The Polish writer Stanislaw Lem frequently had recourse to the gothic trope of the doppelgänger to explore potential temporal paradoxes of space travel (as in The Star Diaries, 2016 [1971]) or unsettling encounters with alien intelligences (Solaris, 2016 [1961]). The 1969 British science fiction film Doppelgänger or Journey to the Far Side of the Sun takes this fictional tendency to an extreme by containing an uncanny double of the entire Earth. But, as I have discussed elsewhere (King, 2022), a series of science fiction authors used twins and, more specifically, the trope of telepathic twins, to explore the ecological understanding of human subjectivity that emerges from the focus of cybernetics research on information exchange. The twin characters in A.E. Van Vogt’s (2013 [1943]) The Weapon Makers, Theodore Sturgeon’s (2000 [1953]) More than Human and Heinlein’s Time for the Stars are bound up, through their telepathic connections, to networked ecologies of cognitive and communicative media. The connectivity between twins is used as a metaphor for understandings of cognition that are distributed across individuals as well as between living and non-living organisms.

The digital twins in 2001: A Space Odyssey are also used to test the limits and blind spots of human intelligence. Matthews examines HAL in the context of research being conducted at mid-century into different approaches to control systems. Unlike other approaches to artificial intelligence that were being developed at the same time and focused on individual tasks and problems, the aim of heuristic machines was to ‘imitate the process of human problem-solving by focusing on what is sufficient rather than rational and logical’ (Matthews, 2020: 252). Heuristics were predicted to play an important part in future space missions. As John Diebold put it in 1964 (quoted in Matthews, 2020: 252),

When the first unmanned spaceship goes to Mars, for instance, no one can predict all the conditions it will meet. But the spaceship’ s heuristic system can be given goals of landing, exploring, and returning, and it will accomplish the mission in its best possible way and adapt itself to whatever conditions it encounters.

According to Matthews (2020: 253), the cause of HAL’s murderous actions ‘is not a simple malfunction or the development of a malign sensibility, but the inhuman application of heuristics’. When its Prediction Centre makes a mistake, HAL sets out to restore the necessary conditions for being able to continue in its foundational ‘conviction’ that it is never wrong. It does so by erasing all traces of its error, including the human witnesses. One of the main ‘messages’ of this section of the film, Matthews claims, is the importance of employing safeguarding measures in the implementation of artificial intelligence systems. ‘There is no method to guarantee the reliability of hardware and the novel demonstrates the need for fault tolerance (detection of error) systems that operate independently of the AI’ (Matthews, 2020: 255).

The digital twin employed by Mission Control is just such a safeguarding measure. As with Apollo 13, it enters the narrative of the film as a response to an error in the AI system. It is intended to facilitate the scrutiny of past decisions and map out their possible future consequences. In other words, HAL’s digital twin also functions to tether the on-board systems of the Discovery One to the ‘terrestrial’ modes of cognition represented by Mission Control. When Dave and Frank find that they have very different perspectives on the reality of the situation from the computer, they turn to Mission Control and their systems of twinning for an epistemic ‘grounding’. In a way that mirrors Mission Control, the Earth-bound spectator attempts to fit the increasingly extraordinary events unfolding on the screen within their own existing cognitive frameworks. However, although the intended aim of the digital twin is to function as a mechanism for epistemic grounding, when error intervenes it can have the opposite effect of untethering terrestrial epistemologies. After examining the output from the digital twin, the usual confidence and certainty of Mission Control is replaced by a ‘skepticism’ and a recognition that the events that are unfolding are ‘rather incredible’. The subsequent severing of communication with Earth and the consequent shut-down of the two-way flow of information needed for a digital twin system marks the beginning of an irrevocable disturbance of epistemic frameworks.

Set loose from the epistemic mooring of the digital twin, and isolated after the deaths of his fellow crew members, Dave has no choice but to abandon himself to the alien intelligence that the Discovery One is being inexorably drawn towards. The trope of doubling reappears during the final psychedelic sections in the film that stand in stark contrast to the highly realistic and detailed focus on technological and socio-technical procedures that characterise the majority of the film. When Dave undergoes a transformation into the ‘Star Child’ (an image of a foetus floating in space, which ends the film) he is repeatedly portrayed as being externalised from himself, watching his own body as it ages. So, whereas HAL was doubled by ‘his’ digital twin in an attempt to contain or mitigate against the unknowability that surrounds space travel, Dave’s doubles mark a dissolution of the knowing subject and a radical opening to the unknown. In this sense, 2001: A Space Odyssey stages the counter-tendencies of digital twin technology as a form of space media that was implicit in the Apollo 13 mission itself. In true digital twin fashion, it did this through a prediction or emulation of possible future directions for space exploration.

Conclusion

In his work on how epistemic proxies carry out the work of ‘standardizing knowledge’ and ‘entrenching’ contingent metrics to the point that they become ‘infrastructural and pregiven’, Mulvin (2021: 6) points out that these ‘stand-ins’ do not function in cultural vacuums. Rather, ‘proxies shape and are shaped by the politics of representation and delegation’ (Mulvin, 2021: 6). To demonstrate this point, Mulvin draws on Murphy’s (2017: 24) descriptions of what she terms ‘phantasmagrams’, that is ‘quantitative practices that are enriched with affect, propagate imaginaries, lure feeling, and hence have supernatural effects in surplus of their rational precepts’. Phantasmagrams, such as the measure of GDP, ‘conjure ineffable realms that can take shape as a collective phantasy in excess of the representational and logical limits of quantification practices themselves’ (Murphy, 2017: 24). Appealing to a long and varied history of using twins as scientific instruments or metaphorical tools, digital twins take on some of the qualities of the phantasmagrams identified by Murphy. On the one hand, the concept of digital twins draws on the use of twinship as a guiding metaphor of cybernetics, rooted in the work of mid-century cyberneticians such as Gray Walter and Norbert Wiener. This in turn draws on the role of twins in the use of genetics research by the eugenics movement. By enabling researchers to differentiate between inherited and environmental factors in human development, twins became an instrument of human control and mastery over nature.

On the other hand, as we have seen, digital twins also evoke cultural imaginaries of the uncanny and that which exceeds human understanding and control. Elsewhere (King, 2022), I have traced an alternative discourse within digital, visual, and literary cultures to the techno-rationalist use of twins that associates twinning with the dynamics of glitching and unpredictable transformation. Rather than predictable and controllable, the future heralded by the ‘glitch twins’ of science fiction and horror is often open, characterised by weird loops that are unpredictable and uncontrollable. When we examine the context of the putative emergence of digital twin technologies as a form of space media in the context of error and the exploration of the limits of human survival and cognitive frameworks, we see these tendencies intersect and overlap. By thinking through their roots in space media, therefore, we can see the potential of digital twins to ‘propagate imaginaries’ beyond the managerial systems of control that constitute their dominant contemporary usages.

This article has been an attempt to prise open the imaginaries of digital twin technologies through an examination of their status as space media. In that respect, I draw on Parks’ (2005) study of satellite technologies, in which she borrows Foucault’s concept of genealogy to ‘open the field of the technological to a broader set of historical and cultural questions that energize new imaginaries rather than restricting our understanding to technical inventions or institutional regulations’ (p. 15). By doing so, Parks (2005: 13) attempts to ‘wrestle the satellite out of the orbit of its “real agencies” so that it can be opened to a wider range of social, cultural, artistic, and activist practices’. I echo Parks by examining the counter-tendencies within the use of digital twin systems in the context of space travel as well as the cultural cross-currents out of which it emerged. By doing so, it is my hope that this article will go some way towards making digital twin technologies available for creative and artistic interventions.