Scott Aaronson's Worldview Manager helps uncover inconsistencies in your beliefs
projects.csail.mit.edu
projects.csail.mit.edu
According to the test there is an inconsistency between:
A simulation is never equivalent to the real thing. (true)
and
One can know that minds exist purely through empirical observation. (true)
The explanation given why these are inconsistent is:
Empirical observation will not be able to distinguish between some being and a simulation of that being, if the simulation is good enough. If simulations can never be equivalent to the real thing, then empirical observation is not enough.
The crux here is that the 'good enough' bit modifies the statement in 'B' after the fact, no simulation is perfect, but now we have to take one that is 'good enough' to be taken for perfect.
Basically, this was an interesting idea, but I don't know if it's possible to write questions that reflect a broad enough range of worldviews to actually categorize them. There are probably hundreds of distinct views of "soul", including both people who believe in them and people who don't.
The explanation appears consistent with a response that says "A simulation may be observed to be equivalent to the real thing (true)".
Perhaps a simple logic slip.
If you are self-aware and you are curious about self-awareness then you can present your environment with stimuli (spoken words, gestures, and so on) to create variations on the Turing test, if the parties appear to think like you do then slowly the evidence over time accumulates to the point where you will have to assume others are sentient.
Testing rocks, cats and people in this respect will give you different ideas about the sentience of rocks, cats and people (and in the case of the cats, if you should frighten them with your sounds or gestures will give you a healthy respect for small creatures with claws).
I believe the people in my environment when they say they have a mind too, I trust cats to have 'some form of mind' because they seem to be self motivated like I am. Their motivations are clearly different than mine, but they do seem to think, at some point the evidence that points in that direction has to be taken to be overwhelming, so therefore we lean towards taking it to be the truth.
No simulation gets in the way of that one.
Now if someone were to slip a simulated intelligence in my environment (we are assuming that it is a fact here for the sake of the argument) and I would not be able to tell the simulation apart from any of the other entities then we have achieved a situation where there would be an equivalence between that simulated intelligence and 'the real thing'.
Maybe I'm stretching things here, but it seems to make sense to me.
Assuming a realist position (on the existence of the universe!) the fact that a cat responds in a similar way that a person with a mind does proves nothing - if I am startled (assuming I have a mind) then it is not the internally thoughtful part of myself that causes me to jump but it is a basic response below the action of my thought processes, I can't decide not to jump.
Consider that androids will soon appear to think - or that if you play poker online against a bot, but think it a person, that this person [bot] appears to you [from your position seeing their moves] as thinking. It is clear that the bot is only apparently thinking it is simply acting according to a complex algorithm.
Taking your point on a "slip[ping] a simulated intelligence" in front of you. You seem to be saying that if the simulation can work then it is a real mind (a sort of pure Turingism). What if a simpler person is convinced that the intelligence is real, but you are not. Does that mean that the intelligence has a mind? Does it think when interacting with that other person but not with you. I contend that it genuinely thinks in neither situation.
[Yes I meant "hear", I have problems with homophones, usually it's their=there and usually I catch it!]
This will become more interesting if it gains the ability to take uncertainty into account more effectively. Unfortunately, I have my doubts about whether their (firmly logic-based) resolution engine can be made to do this without major work.
Broken link, right? I reported it to them a few hours ago and they/he're apparently working on it.
Intentional or not, this is not the kind of message you want to send if you're serious about promoting consistent rational thinking.