by Francois Tremblay (e-mail: FTremblay@liberator.net) [October 9th, 2002]
Various schools of thought have proposed the idea that our world is mere appearance, and that there is some kind of underlying mystical truth that can explain everything.
“[I]t is probable that we are part of the simulations of the human species from the future: we are nothing but a reproduction of the state of Earth as programmed by the real, technologically-advanced humans.”
For example, religious mystics propose that it is the supernatural that is the true reality, meditators propose the absence of thought as a profoundly significant state of being, Idealist philosophers propose a “realm of ideas” which is the true reality, promoters of Near-Death Experiences propose that the NDE is the higher reality, and so on. Any idea or experience which diverges from daily experience is inevitably pointed as the answer.
While it is not part of this extreme, a recent argument by Dr. Nick Bostrom (Department of Philosophy, Yale University) has made modest waves in the media. According to these reports, Bostrom believes that we are in fact probably living in a computer simulation.
His reasoning is fairly simple. There will be a time where we are able to simulate sentient life at a large scale. If that is so, then there will be an enormous number of lives which will be simulated in the future. Eventually, it is not too far-fetched to think that this number will be far greater than the number of people that have ever lived.
True, this argument is based on two premises. First, it would require computational capacities which can scarcely be imagined at the moment, and secondly, it also assumes that artificial intelligence is possible. But we can assume that such capacities will surely exist in the future.
Given that the number of future simulated beings far surpasses the number of living beings, the argument goes, we must conclude that it is probable that we are simulated beings as well. That is, it is probable that we are part of the simulations of the human species from the future : we are nothing but a reproduction of the state of Earth as programmed by the real, technologically-advanced humans.
At first glance, the argument seems statistically convincing. After all, while biological reproduction is limited, computational production of simulated beings is only limited by technology.
But if one looks at the argument more carefully, it is very similar to the epistemic skeptic “brain in a vat” argument.
The “brain in the vat” argument posits a situation where you are a brain in a vat of preserving fluid, plugged with electrodes which transmit information about the world “as if one was in reality”, and adapts to desired movement as well. In such a situation, as goes the argument, one could not make the difference with a person “in reality”, and therefore it is not wise to reject such a situation as a possibility.
Likewise, the “simulated reality” argument posits a situation where we are fooled in thinking that we are “in reality” when we are really in a simulation. While the argument does not seek to attack our understanding of reality, it is still, at its basis, an old skeptical argument in a new garb.
As such, the two realist objections still apply to it. In any such situation, there are two possibilities : either the simulation is imperfect and there is a way to discover its true nature, or the simulation is perfect. If the simulation is imperfect and there is a way to discover its true nature, then we shall eventually find it, and then the hypothesis will be proven. If the simulation is perfect, then it is a redundant hypothesis which is of no rational value, since it cannot be proven or falsified, and explains nothing.
So we have to conclude that even if the “simulated reality” hypothesis is possible, it cannot be considered rationally valid, just like the “brain in a vat” hypothesis.
Actually, there is something else I must tell you. The argument I have refuted is not quite what Bostrom said. Rather, he proposes that there are three possibilities :
- The human species is very likely to go extinct before reaching a “posthuman” stage;
- Any posthuman civilization is extremely unlikely to run a significant number of simulations of their evolutionary history (or variations thereof);
- We are almost certainly living in a computer simulation
Proponents of skeptical arguments would choose the third option as most probable. I have argued in The Hypothesis of Sentient Self-Destructionthat I think the first option is the most probable, and indeed the fact that we do not observe ourselves as part of a simulation so far does seem to confirm my hypothesis.
But whenever I am right on that or not is irrelevant here : unless we have evidence of being in a simulation, the third option is simply badly formulated. Either we live in a computer simulation, or there is no point in even considering it.
I can think of two more objections to the “simulated reality” hypothesis, on theoretical grounds. One fact that seems problematic to me is the fact that we are able to discuss this hypothesis. If this reality is controlled by people running a simulation, then why haven’t they hard-coded obstacles to discussion of the simulation, or at least saw that it was under discussion and stopped Dr. Bostrom’s research ?
My second objectionable fact is the future existence, and destruction, of sentient simulations. I would like to think that a human government that lasts that long would also afford us enough freedom to protect our lives, if anything else. In such a government, surely the destruction of sentience would be illegal (and in case you want to argue that a simulated sentience would be inferior in capacities, we have already assumed that the simulated consciousness was functionally equivalent).
Of course, there is always a last way out : we can simply assume that the computational capacities to make these simulations is impossible, or that artificial intelligence on par with humans is impossible. However, I see no reason to make that assertion, and as we know, it is usually quite foolish to put limits on future technology.
Related articles
