Chapter 4 Robot Evolution and the Fate of Humanity: Pop Culture and Futurology in the Early 20th Century
By the turn of the twentieth century, the pace of technological development had created a more pressing need for considering where such progress might ultimately lead us. During this period, the exploration of potential futures of humanity in a world shared with self-reproducing, evolving machines was attracting a wider audience. Where Butler, Marshall and Eliot had led in considering these ideas in the late nineteenth century, others soon followed. In this chapter we discuss novels, sci-fi and other literature that explored self-reproducing machines in the early twentieth century, and we also cover speculative scientific work from this period. These works demonstrate the wider discussion of such ideas across society, and show that the current popularity of debates about advanced AI and AGI (e.g. (Kurzweil, 2005), (Harari, 2015), (Bostrom, 2014), (Tegmark, 2017)) is actually a continuation of a public conversation that has been in progress for over a hundred years.
4.1 Literary Work
The growing popularity of the dystopian genre in early twentieth century literature was fuelled in part by a fear of how technology might negatively influence the development of human society (Beauchamp, 1986). Here we highlight works from the genre that involved ideas of machine self-reproduction and evolution.
4.1.1 E. M. Forster: The Machine Stops (1909)
The story pictures a future in which humans live underground in personal accommodation where corporeal needs are entirely satisfied by technology (the global, all-nurturing “Machine”). This leaves them free to concentrate on intellectual development, although it also renders them physically degenerate. Forster describes the Machine’s “mending apparatus” that fixes problems and performs self-repair functions, evoking an early image of a machine capable of self-maintenance. It is the collapse of this functionality, brought about by the mending apparatus itself falling into disrepair, that brings the story to an apocalyptic end. Forster refers in passing to the Machine evolving new “food-tubes,” “medicine-tubes,” “music-tubes” and even “nerve-centres,” but these ideas are not explored in detail.
As mentioned earlier (see footnote of Sect. 3.1) Forster acknowledged the influence of Samuel Butler in his work—the vision in The Machine Stops of a future where an increasing dependency upon machines leads to the degeneracy of the human body certainly echoes some of Butler’s concerns (Sect. 3.1). Forster’s image of self-maintaining machines sustaining human life was further developed twenty years later by John Desmond Bernal, whom we discuss in Sect. 4.2.1.
4.1.2 Karel Čapek: R.U.R.: Rossum’s Universal Robots (1920)
Themes of machine (collective) self-reproduction are present in Karel Čapek’s play R.U.R.: Rossum’s Universal Robots, published in 1920 and first performed in 1921 (Čapek, 1920). The robots46 in the play were constructed from biochemical components and designed to resemble humans, but lacked “superfluous” capacities such as feelings or the capacity to reproduce. They were mass-produced in a factory to replace human workers with a cheaper, more productive alternative. The majority of the factory production was carried out by robots themselves, with only the most senior positions filled by humans. However, the complex formula for manufacturing the key “living material” was a closely-guarded secret, recorded by the factory’s founder (Rossum) before his death and kept in a safe to prevent it from falling into the hands of competitors or the robots themselves.
One of the scientists in the factory experiments in making robots with more human-like feelings such as pain and irritability, but this leads to unintended and ultimately disastrous consequences when the robots come to despise their human masters and rise up against them. This, coupled with an unexplained crash in the human birth rate, leads to a stand-off where the robots surround the factory and the people within it, who are now the only surviving humans in the world. The humans realise that their only bargaining chip is the document that explains Rossum’s formula, without which the robots would be unable to produce more of themselves and would therefore die out as a race as members of the current population fall into disrepair.
The climax of the play thus revolves around a struggle for the ownership of the written instructions that would allow the robots to collectively produce more of themselves—a struggle for the ownership of the robot’s DNA, as it were.47 This idea of the collective reproduction of a society of robots reflects some of Butler’s earlier ideas expressed in Erewhon (Sect. 3.1).
4.1.3 Early American Science Fiction (1920s–1950s)
The appearance of American pulp science fiction magazines in the 1920s, and their growing popularity over the decades that followed, provided a medium in which many writers explored the idea of self-reproducing robots and evolving machines.48
Perhaps the first example in this genre was the British writer S. Fowler Wright’s story Automata, published in the American magazine Weird Tales in 1929 (Wright, 1929) (see Fig. 4.1). With echoes of Butler (Sect. 3.1), the story extrapolates the observed accelerating pace of technological development of the time into the far future, to a point when machines no longer rely on humans to service them. The machines become not only self-reproducing, but also able to design their own offspring. This ultimately leads to a complete takeover by machines and the extinction of the human race. The story views the takeover by machines as the inevitable next stage of evolution: similar to Butler’s work, it suggests that the only way this could have been avoided was by “a war sufficiently disastrous to destroy the world’s machinery and the conditions which could produce it” (Wright, 1929, p. 343). As with Butler’s and Eliot’s work before it, the plot of Automata sounds a warning of the unpredictable long-term consequences of machine evolution:
“Even in the early days of the Twentieth Century man had stood in silent adoration around the machines that had self-produced a newspaper or a needle … And at that time they could no more have conceived what was to follow than the first ape that drew the sheltering branches together could foresee the dim magnificence of a cathedral dome.”
— S. Fowler Wright, Automata, 1929 (Wright, 1929, p. 344)
Three years later, in 1932, the influential American sci-fi writer and editor John W. Campbell published The Last Evolution (Campbell, 1932), which also anticipated the eventual replacement of the human race by self-reproducing and self-designing machines.49 However, Campbell’s story is more optimistic than Wright’s, foreseeing a period where humans live in peaceful and co-operative coexistence with intelligent machines, with human creativity complementing machine logic and infallibility. The end of the human race comes not at the hands of the intelligent machines, but when a species from another solar system invades Earth. The invasion prompts the machines to design a new super-intelligent machine to thwart the attack, and this itself spawns further rounds of creation of more sophisticated machines—the final instantiation of which succeeds in repelling the invaders but is ultimately the only surviving species on Earth. Earlier in the story, the last two surviving humans console themselves while contemplating their fate:
“I think … that this is the end … of man … But not the end of evolution. The children of men still live—the machines will go on. Not of man’s flesh, but of a better flesh, a flesh that knows no sickness, and no decay, a flesh that spends no thousands of years in advancing a step in its full evolution, but overnight leaps ahead to new heights.”
John W. Campbell, The Last Evolution, 1932 (Campbell, 1932, p. 419)
Campbell’s vision of a complementary coexistence of humans and intelligent machines is replaced by a darker image in his 1935 story The Machine (written under the pseudonym Don A. Stuart) (Campbell, 1935). In the story a human-like race on a planet named Dwranl50 design a thinking machine that is set the task of making better versions of itself. The outcome is a machine that takes care of all of the race’s basic needs. However, this ultimately leads to the degeneration of the race’s intelligence, civility and its ability to look after itself—a similar fate to those described by Butler in Erewhon (Sect. 3.1), by Eliot in Impressions (Sect. 3.3) and by Forster in The Machine Stops (Sect. 4.1.1). The machine decides that its presence has become detrimental to the planet’s inhabitants, for they are not engaging with it appropriately, but instead treating it like a god; it therefore resolves to leave the planet so that they can learn to live independently once more.
Laurence Manning’s The Call of the Mech-Men (1933) (Manning, 1933) also mirrors ideas first aired by Butler over sixty years earlier. Two explorers discover a group of extraterrestrial robots who have been living in underground caverns on Earth since their spaceship was damaged many tens of thousands of years earlier. The robots are amused when they hear of humankind’s view of itself as master of its technology, remarking (in their stilted English): “Machine gets fed and tended under that belief! Human even builds new machines and improves year by year. Machines evolving with humans doing all work!” (Manning, 1933, p. 381).
Recurring themes of machine evolution and self-reproduction are seen in various stories over the following years. In Robert Moore Williams’ Robots Return (1938), three robots from a faraway planet travel to Earth in search of information about the origins of their ancestors many thousands of years earlier (Williams, 1938). To their surprise, they discover that they were originally designed by humans, and had been sent into space to accompany their creators in escaping a dying Earth. The humans did not survive the mission, but the robots did, settling upon a distant world; there, they reproduced and ultimately evolved into their current state.51 Another tale of robots outliving their designers is Joseph E. Kelleam’s Rust (1939), set on a post-apocalyptic Earth where human-designed robots have survived after humankind has been wiped out (Kelleam, 1939). The robots try to design and build more of their kind before they succumb to erosion, but ultimately fail in their attempts. One further example is A. E. van Vogt’s M 33 in Andromeda (1943), in which a spaceship of human explorers overcome an extraterrestrial intelligence the size of a galaxy by constructing a self-reproducing torpedo-manufacturing machine (Vogt, 1943).
The most explicit exploration of machine self-reproduction and evolution in early science fiction is Philip K. Dick’s Second Variety (1953) (Dick, 1953). The story is set on Earth at the end of a long-running war between East and West, in which Western forces are driven to design killer robots to turn the tide on the battlefield. The robots are highly autonomous, with each generation gaining more sophisticated powers including self-repair and self-manufacture. The robots eventually become too dangerous for the human designers to be near, and are left to reproduce by themselves. Like the machines in Wright’s Automata and Campbell’s The Last Evolution, the robots in Second Variety eventually develop the ability to design their own offspring, and increasingly sophisticated and human-like species of killer robots begin to emerge. Echoes of these earlier stories are also seen when one of the human characters remarks “It makes me wonder if we’re not seeing the beginning of a new species. The new species. Evolution. The race to come after man” (Dick, 1953).
Themes of machine self-repair, self-reproduction and evolution were central to various subsequent works by Dick. Another notable example is his 1955 short story Autofac (Dick, 1955), which ends with a vision of the seeds of self-reproducing manufacturing plants being launched into space.
Other sci-fi works from the 1950s featuring self-replicating machines include Robert Sheckley’s 1955 short story The Necessary Thing (Sheckley, 1955), and Anatoly Dneprov’s 1958 Russian work Крабы идут по острову (Dneprov, 1958) (later published in English in 1968 as Crabs on the Island (Dneprov, 1968)). Dneprov’s story has echoes of Dick’s Second Variety, featuring small self-replicating robots designed as weapons. The robots are set loose on a desert island to compete against each other in an evolutionary arms race to produce ever more effective weapons. The experiment works, but not in the way that the machine’s inventor had envisaged—he is eventually killed by one of the evolved machines (see Fig. 4.2).
Sources of further commentary on themes of robots and computers in early science fiction include (Warrick, 1980) and (Landon, 2008). Further substantial development of these themes was made during the 1960s—we provide references to some of the most distinguished of these later works in Chap. 6.
4.2 Scientific Speculation in the Early 1900s
Beyond these works of literature, the early twentieth century also saw continued speculations from the academic community on the long-term implications of machine self-reproduction and evolution. Here we look at the most notable example of this kind, written by the eminent British scientist J. D. Bernal.52
4.2.1 J. D. Bernal: The World, The Flesh and the Devil (1929)
John Desmond Bernal (1901–1971) conducted pioneering work on structural crystallography, and supervised the PhDs of two future Nobel laureates (Max Perutz and Dorothy Hodgkin) (Holmes, 2006), (Brown, 2005). Later in his career he also became interested in the origins of life (Bernal, 1951). In addition to his experimental science, he authored many works on history, science and society.
Bernal’s first monograph—“The World, the Flesh and the Devil: An Enquiry into the Future of the Three Enemies of the Rational Soul” (Bernal, 1970)—was perhaps his most futuristic writing.53 In it, he discusses how one might examine the far future of humanity in a scientifically defensible way. Bernal begins by sign-posting the methodological and intellectual dangers to be avoided in such an endeavour, and discussing the unavoidable limitations. Keeping these issues in mind, he proceeds to explore what might be said of the three major kinds of struggle facing humanity: against the forces of nature and the laws of physics in general (“the world”); against biological factors including ecology, food, health and disease (“the flesh”); and against psychological factors including desires and fears (“the devil”).
Writing before the advent of space travel, atomic energy or computers, Bernal first tackles how humankind might overcome the challenges that arise from the material world. He argues that limitations of land and energy in the world will eventually compel us to colonise space: “On earth, even if we should use all the solar energy which we received, we should still be wasting all but one two-billionths of the energy that the sun gives out. Consequently, when we have learnt to live on this solar energy and also to emancipate ourselves from the earth’s surface, the possibilities of the spread of humanity will be multiplied accordingly” (Bernal, 1970, p. 22). After discussing plausible technologies for powering a spaceship to escape the Earth’s gravitational field and then to travel in space, he goes on to imagine how humans might set up permanent space colonies.
Bernal proposes a “spherical shell ten miles or so in diameter” (Bernal, 1970, p. 23) which could provide a habitable environment for twenty or thirty thousand inhabitants. After discussing how the construction of a sphere might be bootstrapped from a basic design built largely of materials mined from an asteroid, Bernal continues with a description of the organisation of a mature sphere. It is imagined as “an enormously complicated single-celled plant” (Bernal, 1970, p. 23) with a protective “epidermis,” complete with regenerative mechanisms to protect against meteorites, mechanisms for the capture of meteoric matter to be used as raw material for the growth and propulsion of the sphere, systems for energy production from solar energy, stores for basic goods such as solid oxygen, ice and hydrocarbons, and mechanisms for the production and distribution of food and mechanical energy. The sphere would also have mechanisms for recycling all waste matters, “for it must be remembered that the globe takes the place of the whole earth and not of any part of it, and in the earth nothing can afford to be permanently wasted” (Bernal, 1970, p. 25).
The inhabitants of these globes in space would not be isolated, but would be in wireless communication with other globes and with the Earth. In addition, there would be a constant interchange of people between the globes and the Earth via interplanetary transport vessels. Having set out how the globes might function to sustain life as “mini-earths,” Bernal imagines a yet more ambitious scenario:
“However, the essential positive activity of the globe or colony would be in the development, growth and reproduction of the globe. A globe which was merely a satisfactory way of continuing life indefinitely would barely be more than a reproduction of terrestrial conditions in a more restricted sphere.”
— J. D. Bernal, The World, The Flesh and the Devil, 1929 (Bernal, 1970, p. 27)
Hence, the globe is conceived of as a fully self-maintaining and self-reproducing unit akin to a living organism.54 Bernal discusses various ways in which a globe might construct another globe:
“… either by the crustacean-like development in which a new and better globe could be put together inside the larger one, which could be subsequently broken open and re-absorbed; or, as in the molluscs, by the building out of new sections in a spiral form; or, more probably, by keeping the even simpler form of behavior of the protozoa by the building of a new globe outside the original globe, but in contact with it until it should be in a position to set up an independent existence.”
— J. D. Bernal, The World, The Flesh and the Devil, 1929 (Bernal, 1970, p. 27)
Once the globes have become equipped with the capacity for self-reproduction, Bernal further envisages how an evolutionary pressure to explore might arise among a population of globes:
“As the globes multiplied they would undoubtedly develop very differently according to their construction and to the tendencies of their colonists, and at the same time they would compete increasingly both for the sunlight which kept them alive and for the asteroidal and meteoric matter which enabled them to grow. Sooner or later this pressure … would force some more adventurous colony to set out beyond the bounds of the solar system.”
— J. D. Bernal, The World, The Flesh and the Devil, 1929 (Bernal, 1970, p. 29)
The enormous challenges that would be faced in travelling interstellar distances are addressed, but Bernal argues that such a vision is nevertheless reasonable to consider: “once acclimatized to space living, it is unlikely that man will stop until he has roamed over and colonized most of the sidereal universe, or that even this will be the end. Man will not ultimately be content to be parasitic on the stars but will invade them and organize them for his own purposes” (Bernal, 1970, p. 30).
Moving next to the possibilities of how our own bodies might develop in the distant future, Bernal imagines that humankind will increasingly replace and augment body parts with synthetic alternatives—a vision previously explored by Butler (Sect. 3.1) and H. G. Wells (Sect. 3.4) among others.
Turning to the activities such advanced beings might pursue, Bernal suggests that, among other important scientific questions,55 there would surely be intensive further study of life processes, and the creation of synthetic life. However, “the mere making of life would only be important if we intended to allow it to evolve of itself anew … [however] artificial life would undoubtedly be used as ancillary to human activity and not allowed to evolve freely except for experimental purposes” (Bernal, 1970, p. 45).56
Bernal’s vision of the relationship between the future evolution of humans and machines is more symbiotic than the futures imagined by Forster and Čapek: “Normal man is an evolutionary dead end; mechanical man, apparently a break in organic evolution, is actually more in the true tradition of a further evolution” (Bernal, 1970, p. 42). This perspective is more in line with the ideas expressed by Butler in Lucubratio Ebria (Sect. 3.1), and with those of sci-fi authors such as John W. Campbell (Sect. 4.1.3). Bernal suggests that the main barriers towards progress in these areas would arise from human psychology—in addition to having the desire for progress, we must also “overcome the quite real distaste and hatred which mechanization has already brought into being” (Bernal, 1970, p. 55).57 Various ways in which such barriers may be overcome are suggested, but Bernal does not discount the alternative possibility that we ultimately find ways of living a simpler yet more satisfying life that is not occupied by science or art but more at one with nature.58 He also considers a third possibility—“the most unexpected, but not necessarily the most improbable” (Bernal, 1970, p. 56)—that the future evolution of humanity might diverge, with one race following the natural path and another race following the intellectual and technological path.
Nearly fifty years after the publication of The World, The Flesh and the Devil, Bernal’s idea of a space globe inspired one of the concepts for human space colonies developed during a 1975 NASA Summer Study on Space Settlements (see Fig. 4.3). The concept was named the “Bernal sphere” in honor of his work (Johnson & Holbrow, 1977, p. 48).
4.2.2 The Widening Impact of Ideas
The work described in this chapter, both by sci-fi authors and by scientists, brought the idea of machine self-reproduction and evolution to a wider audience. While scientists paid closer attention to theory and to practical details, that is not to say that the ideas set out in the literary works covered above were not also taken seriously. In particular, Čapek’s play R.U.R. became internationally well known and influential within a few years of its release (Čapek, 2008, p. 9)—Winston Churchill, for example, referred to it in his 1931 essay Fifty Years Hence, in which he discussed what could be said of how the world might develop in the coming decades (Churchill, 1931).59
Another sci-fi author from the period whose work attracted widespread attention was Olaf Stapledon. His 1937 novel Star Maker (Stapledon, 1937) featured an extensive further exploration of the possibilities for humanity’s descendants living on self-sustaining artificial planets, as originally proposed by Bernal.60 Stapledon’s work was admired by figures as diverse as Winston Churchill, Arthur C. Clarke, Freeman Dyson and Virginia Woolf.61
* * *
As we have seen, the early decades of the twentieth century saw an increasingly wide-spread discussion of evo-replicator technology in fictional literature, accompanied by a pioneering, more scientifically-grounded discussion by J. D. Bernal of its potential implications for the far future of humanity. This set the stage for the work of the mid-twentieth century we are about to discuss in the next chapter, which saw a rapid proliferation of contributions from scientists. Along with more detailed theoretical discussions of the potential uses of this technology, the mid-twentieth century also saw the first rigorous work on a design theory for self-reproducing machines, and the first implementations of self-replicator systems in software and in hardware. This period therefore marks the attainment of the third and final step in the history thus far of the intellectual development of thought about self-replicators (Sect. 1.6). It is to these groundbreaking developments that we now turn.
The play introduced the word “robot” into the English language.↩︎
As it turns out in the play, this bargaining strategy of the humans comes to nothing when they realise that Rossum’s document has already been destroyed. The play ends when the last remaining human, Alquist, meets two robots who seem to have developed the capacity for love and human-like reproduction, thereby giving hope that although the human race is about to die out, human-like life will continue.↩︎
Some of the stories mentioned in this section are reprinted in (Asimov et al., 1983). Scans of most of the original publications are also available (for the purposes of private study, scholarship and research) from the Luminist Archives website.↩︎
The name being a near-anagram of Darwin.↩︎
We note in passing that the idea of self-reproducing machines was also touched upon—in a different context—by the respected American biogerontologist Raymond Pearl in his 1922 monograph The Biology of Death (Pearl, 1922). Emphasising that natural selection “makes each part [of an organism] just good enough to get by,” he continues: “The workmanship of evolution, from a mechanical point of view, is extraordinarily like that of the average automobile repair man. If evolution happens to be furnished by variation with fine materials, as in the case of the nervous system, it has no objection to using them, but it is equally ready to use the shoddiest of endoderm provided it will hold together just long enough to get the machine by the reproductive period” (Pearl, 1922, p. 148). This being the case, it is “conceivable that an omnipotent person could have made a much better machine, as a whole, than the human body which evolution has produced, assuming, of course, that he had first learned the trick of making self-regulating and self-reproducing machines, such as living machines are” (Pearl, 1922, p. 148).↩︎
To apply a modern term first introduced by the biologists Humberto Maturana and Francisco Varela to describe living systems, we might describe the globe as an autopoietic organisation (Maturana & Varela, 1972).↩︎
However, Bernal did not see the future intellectual development of humanity resting on science alone: “just as all the branches of science itself are coalescing into a unified world picture, so the human activities of art and attitudes of religion may be fused into one whole action-reaction pattern of man and reality” (Bernal, 1970, p. 54) (a similar idea was later explored by the philosopher John Macmurray (Macmurray, 1956, ch. 10)).↩︎
In this section of the text, Bernal refers to an essay published in 1927 by the Scottish physicist L. L. Whyte, entitled Archimedes, or The Future of Physics (Whyte, 1927). In the context of examining the convergence of the sciences of physics, biology and psychology in the study of life, Whyte discusses approaches to creating synthetic life using chemical components. He argues that higher-level intelligence would have to be evolved rather than designed, and estimates minimum times that might be required to achieve various evolutionary outcomes (ranging up to one million years for mammalian-level intelligence). Whyte—regarded as a maverick by some (Whyte, 2017, p. ix)—proposes the establishment of an International Institute for Evolutionary Research to oversee such a massively long-term synthetic evolutionary study of living systems (Whyte, 1927, pp. 47–65).↩︎
Churchill’s essay also echoed some of the themes raised by Butler (Sect. 3.1), such as the pace of technological development having progressed so far that “Mankind has gone too far to go back, and is moving too fast to stop” (Churchill, 1931). Churchill recognised, as did Bernal before him, that the human mind might struggle to keep up with the rate of development of technology. But in contrast to Bernal (whose work looked over far longer horizons), Churchill thought the solution was not for us to overcome our distaste of mechanization, but to nurture “an equal growth in Mercy, Pity, Peace and Love, [without which] Science herself may destroy all that makes human life majestic and tolerable” (Churchill, 1931).↩︎
However, Stapledon did not explore the possibilities of self-reproduction of the globes, as Bernal had done.↩︎
For details, see Gregory Benford’s foreword in the SF Masterworks edition of Stapledon’s Last and First Men (Stapledon, 1999a), and the quotes at the start of the SF Masterworks edition of Star Maker (Stapledon, 1999b).↩︎