Sunday, August 19, 2007


Many otherwise-intelligent people have an unfortunate tendency to dismiss entire realms of inquiry out of hand. Perhaps the most common example of this is the failure to appreciate the possibility of a priori or non-scientific rational inquiry, i.e. philosophy. The prevalence of ignorant scientism in this thread (bashing Nick Bostrom's simulation argument) is remarkable -- though sadly not atypical.

One commenter suggests that an untestable hypothesis must consequently be classified as either 'myth' or 'garbage'. (He did not tell us how to test this very suggestion. I can only assume he was storytelling.) Another calls Bostrom's argument "pseudoscience gibberish". Yet another chimes in:
This is very much like saying the earth might really be only 3000 years old and $DEVIL just made it seem like its much older to fool everyone.

IOW, it is all hocus pocus claptrap what ifs and doesn’t belong in any science discussion.

The blogger (Peter Woit) himself writes:
I don’t see what the problem is with “lumping Bostrom’s ideas in with religion”. They’re not science and have similar characteristics: grandiose speculation about the nature of the universe which some people enjoy discussing for one reason or another, but that is inherently untestable, and completely divorced from the actual very interesting things that we have learned about the universe through the scientific method.

Really, if people can't tell the difference between a reasoned philosophical argument and random "hocus pocus" or religious proposals... well, let's just say it's further evidence of the urgent need for philosophical education in schools!

If you think that Bostrom's argument is flawed, then by all means put on your philosopher's hat and expose its errors. But this requires actually engaging with the argument. To dismiss it just because it didn't involve any labwork is the worst kind of scientism.

I should add a disclaimer. Sometimes people attack "scientism" when their real target is epistemic standards in general. (See the comments here, for example.) Not me. I'm all in favour of having rationally justified beliefs. What I'm attacking here is the lazy assumption that science is the only source of rational justification. This assumption is simply false (and indeed self-defeating). This should be too obvious for words, but apparently it needs to be said: rigorous philosophical argumentation can also provide rational support for a conclusion.

Hat-tip: Robin Hanson (who offers some incisive criticism of his own).

See also: Explaining Beliefs. (It's the same core issue, really: dogmatic dismissal is no replacement for reasoned inquiry. You can't tell whether a question is answerable until you try.)


  1. When I attack scientism, I'm attacking the same thing you are: the misguided and self-defeating belief that science is the only grounds for truth, and specifically the only source for "rationally justified belief."

    It's true, I tend to be suspicious of epistemic standards in general, not because I think that they're inherently flawed (as scientism is), or because I don't think we should apply and attempt to live up to them. I do that myself. I'm suspicious of them only because I know that any epistemicth standards worth applying are impossible to live up to.

  2. the simulation argument doesnt get you out of the dooms day argument. given we are in a simulation we will be 'deleeted' in the next "X" years based on the standard doom's day argument.

    Furthermore the indifference principle also then implies somthign about the civilization running these simulations - ie that the civilizations cannot be older than a certain age otherwise the chance of them devoting huge amounts of computing power to this part of the simulation and not more of it becomes diminishingly small.

    So you would get a limit for REAL human civilization (the one that is simulting us) that isn't all that much further out than the one for this simulation.

    So the question then starts to become interested in what sort of computing power we can accumulate in the next few thousand years. we don't have to worry about the fraction of that resource likely to be devoted to ancestor civilizations because thats taken into account by the indifference principle already - although it might be that we take less programming space than a future human - it is debatable if that helps either (cf my analysis

  3. I think Husserl in 'The Crisis of European Sciences',said something like - the problem with empirical science is not that it is not successful, but that it is too successful.
    People blindfolded by empirical science' success seems to get uncritical about the boundaries of science.

  4. Arguing on the general net has got to be less conductive to reasoned philosphical debate and more to testible standards. We do not share philosophical grounds, we do share the testible world.

  5. This is a paper I'll try to read some day when I have the time, looks like it has some fun stuff to play with, but on the face of it you picked a bad exemplar of non-scientific reasons. If you had chosen to attack scientistic attacks on ethical realism or free will or qualia or abstract objects I'd understand, but it's not clear that there's a credible defense to be made of his claim.

    As a side note, the fact that he has an argument doesn't put his claim above religious claims: theologians make arguments for their position too. Crap arguments, maybe, but they're still arguments after a fashion.

  6. "This should be too obvious for words, but apparently it needs to be said: rigorous philosophical argumentation can also provide rational support for a conclusion."

    rational yes, factual depends on whether you confront it with that old thing called reality.

  7. I should clarify that I'm not endorsing Bostrom's argument (though it could be interesting to look at some other time), but merely criticizing a particular way of rejecting it.

    Hallq - scientism is idiotic regardless of the worthiness of its target. Bostrom's argument may be a bad one, but there's no way they can tell that until they actually bother to engage with it.

    Insofar as "theologians make arguments", they are engaging in philosophy of religion. The "religion" of our everyday acquaintance is rather different -- just people asserting their "faith" without any attempt at rational support. The comments I quote are clearly suggesting that all non-science is equally baseless.

    Anonymous (please choose a unique pseudonym as per my commenting policy) - rational support just is evidence of truth (hence, factuality). Unless you think there's a difference between what's "factual" and what's "true"?

  8. > it could be interesting to look at

    I do believe I have fatally wounded it already, I doubt there is much more to look at!

  9. From a different angle, if the simulation included us, but not the environment, then it is somewhat irrelevant whether I'm a computer simulation or neuronal netware.

    If the environment is part of the simulation, I think the amount of computing power to give a non-contradictory environment is greatly underestimated.

    True, all that has to be provided is what passes through our sensory links. But I think contradictions would show up quickly enough unless the simulation obeys all the laws of physics.

    The operator detecting contradictions, pausing, erasing, and rerunning may end up taking even more computation power.

    I think detection of contradictory experiences in the minds of the simulation itself might be computtaionally intensive enough and growing exponentially with the number of minds in simulation; and the control of the simulation itself might be overwhelming. Bostrom has made no attempt to estimate the computation powers of these.


    I should note that our computational powers are **so** **impressive** that having to type in 'edomvogh' copied in from a slightly distorted display is sufficent to distinguish a human from a spam bot.

    Whomever is estimating the quality of computing power available is grossly overestimating it.

  10. >If the environment is part of the
    >simulation, I think the amount of
    >computing power to give a
    >non-contradictory environment is
    >greatly underestimated.

    Whenever this discussion arises, people seem to be assuming that it is humans running the simulation, in real time, in something similar to our universe. Why? In the same way as I can make a simulation of objects in 1-D space, beings in some 100-D space could run a simulation of a 3-D world at arbitrary speed... our perception of time is related to the duration of things in our world (e.g. how long it takes to read this parenthesis.) If the simulation is sped up or slowed down, we wouldn't notice.

    I think a lot of people dismiss the simulation argument because it can be put in the "anything is possible"-box along with parallel universe theories etc. It is simply too 'out there' and non-practical to care about.

  11. [Discussion of Strong AI shifted to here.]

  12. "I'm attacking here is the lazy assumption that science is the only source of rational justification."

    Rightly so. But empirical facts (established by science) have precedence over merely logical inference (like in philosophy). I'd say that non-contradictory speculation is the primary domain of philosophy, clearing a path for science. So rightly understood, philosophy, if logical and coherent, is proto-science.

  13. I always wonder, what would happen if we couldn't label it "science"?

    Would there be a "this versus that" ... or a "scientism"? Or, would we just discuss ideas and draw logical conclusions based on evidence, to be supplanted when better evidence can describe those conclusions?

    Alexander is wrong with the typical "established by science" line of thinking. Empirical facts that Alexander describes were established by people working together in the process ... not by the institution itself.

    To me, that's where this label "scientism" comes from ... when "science" tries to establish some sort of dominance because of a process that people follow to explain our world. There should be no dominance, no war, no "this versus that."

    Instead, lets just use all available tools at our disposal to explain the Universe we live in. If you want to call one of those tools "science" then so be it. Just don't let "science" become more than that (a tool).


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)