Thursday, March 8, 2012

Emergent Humanism

Sim City

A Response to Are You Living in a Computer Simulation?
          by Nick Bostrom (2003)


We are living in a densely textured computer simulation, a highly calibrated imitation of human reality. The world we believe to be real, and our selves within it, are the fabrications of a higher, synthetic intelligence. (“Life is just a dream you know/ That’s never ending,” Cowboy Beebop Blue).

Socrates. I shouldn’t be at all surprised if Euripides were right to say
Who knows if all this life of ours is death
And death is life?
Plato. Gorgias.

Someone has already – according to a probabilities scenario based on the calculations of philosopher Nick Bostrom – implemented the simulation in which we are living. That someone is a posthuman member of a future generation who uses a super-powerful computer to run detailed simulations of his forebears. The posthumans of the simulation argument are oddly nostalgic for the human world. Indeed, the posthuman running our simulation takes man as the measure of his verisimilitude project, going to great(ly extrapolated) lengths to create an “ancestor simulation” of human life on planet Earth.


What is a posthuman? A posthuman is “a radically enhanced human” (Bostrom, Transhumanist Values, 2002, p. 2) In one iteration of Bostrom’s vision of the far future, posthumanity constitutes “a radical change in the human condition” that may be brought about by biological enhancement (genetic engineering, e.g.) or other causes (Bostrom, The Future of Humanity, 2007, p. 20). If realized, posthumanity would engender levels of economic and technological development far surpassing the productivity and ingenuity of humans. Posthumans would enjoy vastly increased life spans, control over pain and suffering, and populations in which a large fraction of citizens have “cognitive capacities more than two standard deviations above the current human maximum.” (Future of Humanity, pp. 20-21). Superintelligence, the signature feature of posthuman civilization, will likely come about through access to AI in computers or computer networks, computer/ human interfaces, or biological improvement of human cognition. (Future of Humanity, p. 22)

Yet having predicted the likelihood of this brave new world and then envisioned it, Bostrom imagines in his simulation argument that members of future generations who run ancestor simulations will put the human at the center of the simulation, orienting the simulations to human perception.


Man is the measure of all things (“of white and heavy and light and everything of that sort,” Plato. Theaetetus) in the posthuman simulation. Every percept within the simulation is calibrated to match the human sensory apparatus and to delude it. It is as though some hyper-intelligent demiurge had placed man, himself a simulacrum, at the center of a synthetic universe. Bostrom’s hypothesis curiously mingles probabilities analysis with mythical-religious concepts of a higher power, AI iterations becoming minor deities. In positing the simulation’s multilevel reality, Bostrom draws “some loose analogies with religious conceptions of the world” in which the posthumans running our simulation are “like gods in relation to the people inhabiting the simulation.” (p. 22)


A significant component of Bostrom’s argument for a posthuman simulation consists in detailing how the simulation could be implemented in terms of human perception by the hypothetically advanced computers of the future now representing our present, how in effect posthuman simulators could stage the complete illusion of contemporary human society backed by human history.

Design is a key feature of the simulation: “The microscopic structure of the inside of the Earth can be safely omitted…On the surface of Earth, macroscopic objects in inhabited areas may need to be continuously simulated, but microscopic phenomena could likely be filled in ad hoc.” Filled in: as by a graphic artist or animator. As for the simulated humans in the picture: “(But) in order to get a realistic simulation of human experience, much less {than a simulation of the entire universe down to the quantum level} is needed – only whatever is required to ensure that the simulated humans…don’t notice any irregularities.” The simulated humans – that’s us. We’re inside an unreal reality like the one in The Truman Show or Pleasantville, only we too are simulated.


Underlying Bostrom’s emphasis on the implementation of computer simulation through the channel of human perception is the premise that human knowing comes primarily through the senses. If the posthumans do a good job at mimesis then, the homunculi inhabiting the posthuman-created universe can be duped into believing that it, and they, are real. But how, indeed, will posthuman simulators go about designing the human construct? If the AI makes use of human DNA, might human knowing beyond sensory perception imbue the human model with dreams, archetypes, memories, with
--> déjà vu, intuition, and ESP, perhaps with a template, preternatural, for Beauty, for Justice, for Truth? And with that template – what existential unease, what cognitive dissonance, in the homunculus with human DNA, computed into life within a hyper-reality.

What capacity for memory, then, might animate the simulated human being? As Bostrom supposes, “There is also the possibility of simulators abridging the mental lives of simulated beings and giving them false memories of the sort of experiences they would typically have had during the omitted interval.” (p.23) We know that, already, scientists can abridge memories in lab rats (Todd Sacktor, World Science Festival program on "The Unbearable Lightness of Memory," 2011) In humans though, how many instances of abridging individual memories might it take to alter the capacity to form episodic memories? Would such tampering have the potential incrementally to diminish the credibility of the human memory function? If the simulated human's capacity for knowing through memory were significantly compromised, would these still be humans who were being simulated? For simulated humans with altered or diminished integrity of memory would have been deprived of a crucial part of their humanity.

Socrates. That, then, was the drift of my question, what terms should be used to describe the arithmetician who sets about counting or the literate person who sets about reading; because it seemed as if, in such a case, the man was setting about learning again from himself what he already knew.
Plato. Theaetetus.


Socrates’ inquiry did not encompass topics such as AI and singularity, and it would be mistaken to expect the understanding of the thinkers of classical antiquity to reach as far as the 21st century in matters specifically related to changes in society brought about by technology. It may be useful, however, to consider the ways in which the speculative reasoning of contemporary futurists reflects paradigms and tropes that go back to Plato and the Presocratics. Is the posthuman simulation an iteration of Solipsism? Does the godlike simulator resemble a Platonic demiurge? Has Bostrom, with his focus on the primacy of perception and man as the measure, imagined a futuristic simulation based on uncertain premises?


“Moreover, a posthuman simulator would have enough computing power to keep track of the detailed belief-states in all human brains at all times.” (Bostrom, p. 9) Enough computing power may be, but how can this futuristic computer access and monitor belief-states unless it has incorporated the human genome that will enable it to understand belief? Supposing then (for supposition is the game here) that the posthuman simulator has incorporated human DNA, might not the human template for ethical behavior become part of its computational values?

Bostrom takes another route to equipping his system with ethics, elaborating his analogy with religious conceptions by positing moral guidance within the posthuman simulator in the form of “a naturalistic theogony.” (p. 22) Here the structure of the simulation as perceived by its inhabitants – imagined as a multilevel architecture with a hierarchy controlling it bottom-up from the “basement-level” of reality – provides the scaffolding for ethical behavior. In Bostrom’s posthuman utopia, “all the demigods… are subject to sanctions by the more powerful gods living at lower levels.” (p. 22) Ethical behavior would be promulgated by “the constraints imposed on its inhabitants by the possibility that their actions on their own level may effect the treatment they receive from dwellers of deeper levels.” (p. 22)


At each level of the simulacrum a member of a theocratic hierarchy implicitly directs moral behavior. Does not this utopian high-rise bypass the State, however? What model of the State should we expect to encounter within the posthuman simulation?

Socrates. We said that the boldest would not go to the length of contending
that whatever a State may believe and declare to be advantageous
for itself is in fact advantageous for so long as it is declared to be so.
Plato. Theaetetus.

In the process of fleshing out the simulation argument, Bostrom makes two plausible assumptions: 1.) The posthuman simulator, or “director”, can delete memories in the human brain, rewinding and fast-forwarding like a film editor. (p. 9) 2.) Posthumann simulations will be run by “relatively wealthy individuals who desire to run ancestor-simulations and are free to do so.” (p. 25) Bostrom’s first assumption envisions the posthuman simulator, the “director”, as a highly skilled technician. Bostrom’s second assumption implies a Libertarian conception of government, whose role it is to allow – and to refrain from regulating – the activity of the enterprising individual. This Libertarian strand of the simulation argument prompts certain specific concerns:

· Personhood. For the purposes of running ancestor simulations, might not corporations and other entities be considered posthuman persons entitled to run simulations? (This is not especially far-fetched in light of a recent decision by the U.S. Supreme Court defining corporations as “persons” entitled to free speech rights as donors to political campaigns.)

· Motivation. Why might wealthy individuals desire to run ancestor simulations? Motivations might range from entertainment and commerce at one end of the spectrum, to the study of anthropology, human biology, ethology, evolution and history at the other. Should the simulation argument include the supposition that any of these simulation projects are regulated, conducted, or prohibited by the State?

· Ethics. What oversight of the posthuman individual/ corporation/ government running the simulation might be instituted? Such oversight would be in addition and external to any oversight provided by the theocratic hierarchy within the simulation.

One way humanists might further the dialogue on posthuman simulation is by raising, in this context, the type of questions about the role of the State that they are skilled at formulating.


For the Socrates of Plato’s Theaetetus, the future meant little more than the immanent, testable future of the individual (“Is it also true…that {man} possesses within himself the test of what is going to be in the future?”) By the 20th century, futurists had begun to envision humans, or posthumans, hundreds of thousands of years hence.

Today in the early 21st century, thinking about deep time is still the prerogative of science and science fiction. On the whole, humanists have preferred the past to the future, classicism to futurism, and human history to evolution. Humanist writer and mathematician Blaise Pascal (1623-1662) contemplated man poised between the double infinities, a before-and-after of human existence in which the prior infinity becomes a nothingness.

For, finally, what is man in nature? He is nothing in comparison with the infinite, and everything in comparison with nothingness, a middle term between all and nothing. He is infinitely severed from comprehending the extremes; the end of things and their principle are for him invincibly hidden in an impenetrable secret; he is equally incapable of seeing the nothingness from which he arises and the infinity into which he is engulfed.
–– Pascal. Pensées.

Humanism, with its historical roots in Christianity as well as classical antiquity, has been discouraged from thinking very far into humanity’s future by the Christian doctrine of the end of time, as well as by religion’s ambivalence toward science.


Transhumanism is an emergent form of Humanism. By placing human civilizations within the perspective afforded by deep time, transhumanism calls on us to think critically about humanity's future. The hypothesis formulated in the simulation argument has inspired philosophers, social scientists and scientists to think about humanity's ultimate destiny in terms of possible extinction, or transition to a posthuman state of being. Additionally from the perspective suggested by deep time, transhumanism promotes inquiry into the human condition, and a potentially rigorous testing of human limitations. Humanists working in the humanities would do well to accept the challenge posed by transhumanist inquiry, bringing to the colloquy the tools of postmodern and modern analysis as well as those of traditional exegesis.


Are You Living in a Computer Simulation? was written in 2003, at a time when it was reasonable to assume that any emergent posthuman AGI would owe its existence to the development of non-biological machine learning systems. Nick Bostrom’s modification of the simulation argument in A Patch on the Simulation Argument (2011) doesn’t alter this assumption in a significant way. The surge in research and development in Neuroscience from 1995 to the present, however, has increased the likelihood that AGI will be built through Systems Neuroscience: a hybrid approach combining machine learning systems with brain research (Hassabis, Demis, Winter Intelligence Conference, Future of Humanity Institute, 2011). What difference does this transformation in AI research make to the simulation argument?

Are You Living in a Systems Neuroscience Simulation?

What might Neuroscience have to contribute to the posthuman simulation argument?

According to cognitive neuroscientist Demis Hassabis, there are already adequate machine learning programs for building two of the human brain’s core capacities for knowledge, symbolization and perception, into AGI. What is missing in the scientific schema for building AGI are algorithms for the third core capacity for human knowledge: concept acquisition and representation. The conceptual ability of the human brain, Hassabis suggests, can best be built into AGI using discoveries in Neuroscience regarding concept formation. Neuroscience has opened a door to understanding, and implementing, human knowledge contained within a complex system linking symbol formation to abstract ideas, concepts to perception, knowledge to memory, and the conscious mind to the unconscious through the translatable language of dreams.

If a posthuman simulator exists, it would have a massively enhanced, but at its core human, brain. The posthuman simulator, as demiurge, would then be subject not only to ethical constraints built into the architecture of its ultra-intelligent machine learning system, it would be imbued with the capacity for moral judgment derived from the conceptual reasoning capacity of the human mind. Like the computer scientists of the 21st century, the posthuman simulator would be aware that the simulated human not only possesses complex apparatus for perceiving the world around him, but a mind capable of reason and judgment. Any simulation of reality would therefore have to take into account human knowing through memory, dreams, and abstract reasoning as well as through observation and perception. It would not be enough for the posthuman simulator to create a realistic imitation world, for the simulated humans within it would understand that world with their human minds. The presence of concept formation in the human model would mean that high-level abstractions like Beauty, Justice, and Truth could be linked through digitized neural pathways to the physical substrate of the simulated brain. Possessing Truth as a core value, the human would be driven to protect his core, and to validate his surroundings, much as a machine learning AI system would be driven to preserve its core utility functions and to clarify its goals (Omohundro, Stephen, The Nature of Self Improving Artificial Intelligence, 2007).

Moreover, as Hassabis’ research on short-term episodic memory suggests, any “abridging” of discrepant or anomalous memories would be vastly complicated by the way human perception exists in a continual feedback loop with concept formation, memory, and the reconfiguration of memory through dreams. Cognitive science research demonstrates how emotionally heightened episodes are culled from among more mundane memories and replayed as dreams. The neurological pathway from event to memory to dreaming to knowledge produces a survival advantage, and a possible long-term change in understanding. As imagined in dramatic poetry, dreams made of memories may eventuate in self-knowledge, perhaps in transpersonal wisdom:

Even in our sleep, pain which cannot forget
falls drop by drop upon the heart
until, in our own despair, against our will,
comes wisdom through the mind of Zeus.

Agamemnon. Aeschylus. (l. 176-181)

The “pain which cannot forget” is encapsulated in an emotionally heightened memory. In Aeschylus’ poetic understanding, sleep and dreams convert pathos – experience or suffering – into emotional intelligence, or wisdom, over time (“drop by drop upon the heart”). This is an unconscious, non-intentional process (“against our will”). In Aeschylus, the wisdom that ensues comes from a mind more powerful and capacious than the mind of man. To the ancient Greeks, the mind of Zeus was a meta-mind, Zeus having formed unions with Metis (intelligence), Themis (the laws of the physical universe), Eurynome (the laws of society), and Mnemosyne (memory) before establishing his reign on Olympus, as related in Hesiod's Theogony (Bulloch, Anthony, Classical Mythology, 2012).

For the purpose of illustrating the simulation argument, which owes its credibility to the rigor of probabilities, its author, Nick Bostrom, considers questions relating to philosophy of mind peripheral to his essential argument. As philosophy of mind bears directly upon the question: What is a human? and therefore: What is a posthuman? however, its concerns should not be marginalized. Is there at present a human nature, and is wisdom part of it? Can the mind of a wise man be uploaded (the wisdom of the Buddha? the sense of humor of the Dalai Lama?) or do some qualities of human mind elude even the intricate mapping of Neuroscience?

The simulation argument is a provocative hypothesis that prompts us to think in new, more scientifically based and rigorous ways about humanity’s future. As the idea of the posthuman simulation continues to be tested by critique and discussion, and re-imagined by those who are inspired or disturbed by the scenario it presents, it will hopefully be the case that more and more people become less and less afraid of technological change in the 21st century and beyond. For “the future comes apace; what shall defend the interim?” Technological change can take us to many places, but never back.


Anthony Bulloch said...

[Long post in two parts - please read both, as they belong together.]

Part I:

Hmm - I have to admit that my mind/brain just fuses (blows a fuse, too) when I try to get it round a topic like this. Patently a serious deficiency, or disability, of my mental/intellectual equipment, because it's the same when I read Plato - I read, then re-read, then re-re-read Socrates presumably wise remark in response to the stupid or mundane Euthyphro (or whoever) and then sit wrestling with a total blank on my mental screen - 'huh?' is all I can come up with, no matter how much I try to gather something in. I'm in that cave trying to catch one or two fluttering birds, but there aren't any, to be seen or to be heard. So.... is this like Romney's Etch-A-Sketch? Only the etch-a-sketch was actually a momentary apparition on a post human-Romney's Etch-A-Sketch...

[.....go on to Part II which follows in the next post.]

Anthony Bulloch said...

[This is part II of a two-part post: please read Part I first.]

Part II:

One thought, then: how do we know that the posthuman whose world we have been imagined into isn't him/herself a product (or would it be 'function') of a 'higher' post human, who is him/herself a figment.... but you get the idea. Why not an infinite regress of post humans? If there were/is such an endless regress of worlds, does it, would it, matter? How would we know that it does/would? And what of our moral sense? Would that even be possible in such an Etch-A-Sketch world? Suppose each level of posthuman operators - each one behind the other behind the other and so on in perpetuum - erased their Etch-A-Sketch all at the very same time? 'Would you say, O Euthyphro, that Being would be something? or would it be nothing? And what about Morality....?'

linda colman said...

You are not alone in your experience of feeling the mind boggle at theoretical constructs like the simulation argument. The scenario you paint involving an infinite regress of posthuman curators is actually not dissimilar to Bostrom's multi-level hypothesis, in which the posthuman curators are 'encouraged' to behave morally by knowing their place within a complex hierarchy. All the questions you raise are good ones, but rather than address them as such, I would like to point out a feature of the simulation argument that can easily get lost amidst all the speculative thinking. The author of the simulation hypothesis, Nick Bostrom, does engage in speculative thinking, like Socrates. However, there is a significant way in which the simulation hypothesis differs from, say, Zhuangzi's dream about the butterfly.(continued)

linda colman said...

The key distinction is this: Bostrom's argumentation, unlike that of a Socrates or a Zhuangzi, is based on probabilities. He lays out the probabilities core of the simulation argument in the original paper, for which the URL is provided at the beginning of this essay, and in a follow-up entitled "A Patch for the Simulation Argument." The 'patch' uses probabilities to explore the likelihood of 3 possible outcomes for the human race, only one of which is: "you are living in a computer simulation." One of the other propoitions explored is that the human race will become extinct before it reaches a posthuman stage. Now that really makes the mind boggle: total extinction. It is possible, then, that while Bostrom no doubt takes the sim hypothesis entirely seriously, he has also used it, imaginative as it is, to invite thinking about more difficult issues such as global catastrophic risk. (continued)

linda colman said...

Bostrom's paper, written in 2003, has received a wide-ranging response, including attempts by serious academic philosophers like John Searle to come to terms with its premises. In the meanwhile, Bostrom has made a significant contribution to thinking about artificial intelligence, the Singularity, and transhumanist thought. Bostrom is, in addition, the founder and director of the Future of Humanity Institute affiliated with Oxford; the institute is in the vanguard of conducting studies on events affecting humanity, but in the long to very long term: enhancement, identity, AGI, global catastrophe. Few other research entities have the collective brainpower or the bucks to conduct inquiries of this kind; thus the FHI is a valuable resource. It is also, in a very real sense, where the action is. Yet many in the humanities remain blissfully unaware of such enterprise. It is time to move beyond Fukuyama's hysterical response to Transhumanism. It is time for those in the humanities to engage with the issues raised by futurist thinkers. So let's un-boggle our minds, and connect with the new ideas about humanity.