Lately I've been reading some transhumanist science fiction and watched some futurist shows with transhumanist elements on YouTube. I've enjoyed both immensely, but for reference I'd like to spell out why I think they are about as grounded in reality as your average dragons-and-sorcery fantasy.
Transhumanism is basically the belief that the human race can evolve with the help of science and technology. That belief in itself is ill-defined enough that there is nothing there to argue against - after all, thanks to vaccines we don't die as much as we used to, and thanks to dental care we keep our teeth for much longer than we used to - and if you just define "evolve" suitably you can prove anything.
1. Computational Equivalence
The claim that I have an issue with is more specific: it is the idea that computational equivalence means that human minds and human consciousness can run on any Turing-complete computer, and in particular (given enough memory) on a computer like the one you are reading this on.
My counterargument is that the usual definition of computational equivalence is not sufficient for this case. Basically, it states that if you have a computation C that for every x as input produces a given y as output, and if you create a computation C' that for every x as input produces the same y as output as C , then C and C' are equivalent. What this boils down to is that if you can transform the input from the input expected by C to the input expected by C' , and the transform the output of C' into the same format as C produced, then you have equivalence and you can treat C' as C for all intents and purposes.
In particular, just because you and a simulation of you produces the same answer to a question, it doesn't follow that the simulation is conscious.
2. The Golem
This is where I think the transhumanists follow the same path as the renaissance Jews that imagined golems. A golem[c] is an artificial human, made of clay or dust, and animated by providing it with a holy Word, inscribed on its forehead or given to it on a piece of paper. Through the Word, God's animating power would enter the golem, and it would live. Remove the Word, and it would die.
Since science and biology was still in the future, renaissance Jews did not understand what caused animals to live and to move - so they came up with their own equivalence: Since God created Adam from dust, and gave him life by breathing the breath of life into his nostrils, then creating a body of clay or dust, and giving it the force of the divine Word, would surely result in a living thing. They didn't understand biology, but they did understand scripture; so that was what they used to construct their stories about the world.
The transhumanists claim that the mind is software, and since it doesn't matter what substrate you run the software on, "running" a human mind on a computer is the same as running it on a human brain. However, we don't know if the human brain is computationally equivalent to a computer. We don't know what the requirements are for a substrate to not just yield the same outputs, but to actually yield a thinking mind that experiences things. It's just that we live in a society dominated by computers, and just as psychologists attempted to explain the human mind with "pressures" that "built up" and had to be "released" back in the 19th century when steam power was dominant, we use what we have to build a framework for comprehending the world around us.
3. A Thought Experiment
Let's consider the following statements to be true:
For a computation, it doesn't matter if an instruction is run on a CPU over here and the next on a CPU over there. All that matters is that both instructions run in as-if-sequential order.
Real-time doesn't matter. Minds can be paused just like your laptop can be hibernated. We could, in theory, hibernate for millions of years between instructions.
A mind exists even if we never find out about the outputs of its substrate. For example, if you had a black box with an artificial mind in it, and turned it on for a minute, even if you could never observe any effect, the mind was there, and it was thinking, for that minute.
Now consider the fact that the NAND instruction is universal. That is, it is possible to create a program from nothing but NAND.
NAND can be implemented using billiard balls[e]. Air molecules are like little billiard balls. Therefore, as you read this, there are the equivalent of computations happening all around you in the air you breathe. Those computations simulate a mind.
Now, you could argue that the pattern of air molecules are chaotic, and that while we can find a set of molecules here and there that by random chance happen to line up in such a way to perform the dance that a billiard-ball NAND gate would do, there would be no connection between those instances.
But you have, in accepting premise (1), accepted that this doesn't matter. If you want to go back on that, you need to define a constraint on the substrate that requires the computation to operate on the same physical objects from instruction to instruction.
Then maybe you argue that these instructions happen at random intervals, or in the wrong order.
But you have, in accepting premise (2), accepted that this doesn't matter. As long as the instructions follow each other according to some ordering that can have arbitrarily large gaps, it doesn't matter. If you want to go back on that, you need to define a constraint on the substrate that requires the computation to operate in hard real-time.
Finally, you may argue that this is all pointless. We can't contact this air-mind. We can't even point to where it is.
But you have, in accepting premise (3), accepted that this doesn't matter. It exists somewhere.
Finally, you may argue that, since this reasoning can be applied to any gas, this implies that your farts are conscious - and then you'd be well on your way to becoming an Internet Transhumanist.
|There is some irony here.|