[extropy-chat] Scientific standards of evidence

[Eliezer S. Yudkowsky]

Chris Phoenix wrote:
> 
> The final straw came a few nights ago, when I remembered something I'd
> been trying for a year to forget.  One evening last summer, my wife
> suddenly got a pain in her lower ribs, in a place she'd never had pain
> before, bad enough that we canceled our plans to go for a walk so I
> could rub her back (and this had never happened before either).  When I
> started the massage, she said, "How come the more you touch me there,
> the more I want to cry?"  A few minutes later, the pain went away.  And
> a few hours after that, we learned that her brother had been killed in a
> motorcycle accident within a few minutes of that time, and his lower
> ribs had been run over.
> 
> Obviously this is not scientific evidence, it's not a repeatable
> phenomenon, and it leads to no useful theory.  I don't ask anyone else
> to accept it or act on it; that's not the point.  The point is: Is it
> more "scientific" for me to ignore it, or to accept that it happened and
> I can't explain it and it might be significant?  Any scientist worth
> their salt would tell me that, since there's no such thing as ESP, there
> must be some mundane explanation.  I told myself that for a year.  But
> I'm becoming convinced that this is the wrong answer.  It's the answer
> that makes scientists see only what they expect to see, every time. 
> Science cannot progress if it can't deal with the unexpected, and
> science without progress is dead.  But modern science, it seems, has
> only two categories: Things that can be studied with the scientific
> method, and things that cannot be addressed.  There is no category
> anymore for observations that cannot be categorized, but only
> catalogued.  If there is such a thing as a mundane theory that makes ESP
> possible, science will find it only by chance.  It's heresy even to
> look.  

I don't ask anyone not to catalog these kinds of incidents, or not to 
investigate them, or not to try and repeat them under experimental 
controls.  Nonetheless, I am willing to assert flatly that nothing 
paranormal happened.

What is the ordinary-world explanation?  Well, for all I know, you are 
posting this story as a test to see if anyone tries to explain it away, 
give it a pseudo-rational explanation when in fact you just made it up and 
there *is* no explanation for why that sort of thing would happen under 
the laws of physics, because it didn't.  I do not need to try and give 
your story an ordinary-world explanation; even if I can't think of any 
ordinary-world explanation at all, I am nonetheless confident enough in my 
understanding of the universe to not feel discomfited.  This world is my 
home, and I know it now, and I know the world doesn't work that way. 
There are a few things left that I don't know yet, but the remaining 
uncertainty does not have enough slack in it to permit ESP.  This world 
could easily be a computer simulation, but if so it is a simulation of a 
world without ESP.

Supposing that I try to think of an "ordinary" explanation for your story, 
what springs to mind?  Several things.  I don't know if the name 
"Elizabeth Loftus" rings any bells; but human memory is far, far more 
pliable than people like to think.  It is possible for a researcher, by 
asking leading questions, to create a memory completely out of fabric - 
for example, of being lost in the mall as a child - and later the person 
will not remember that it is a false memory.  Did your wife really have a 
pain in her lower ribs, or her back?  Did you learn about her brother's 
death a few hours later, or a day?  "How come the more you touch me there, 
the more I want to cry?" sounds to me like not at all the sort of thing 
that is said spontaneously, but very easily the sort of thing that might 
be "recalled" afterward.  It's a scary thing, but no, you would *not* know 
if the memory was wildly distorted - that is the message of the literature 
on memory pliability.  In effect, it seems that each time a memory is 
recalled, it is written back to long-term memory anew.  Then there is the 
selection bias, applied to you personally:  You have an entire life to 
select from to try and recall strange events.  There is the selection bias 
from my perspective; one person on a mailing list has one strange moment 
in their lives, but that's what repeated.  Is the horrifying malleability 
of human memory, combined with selection biases, enough to explain the 
incident as you have reported it?  Possibly, though it still strikes me as 
a bit of a stretch.

Regardless, people who have studied the cognitive science of human error 
tend to be much more wary of people's memories - not their honesty, mind 
you, but their honestly reported memories - than non-CSers.  I expect that 
plenty of people on Earth have memories of events for which there is no 
rational explanation, because they never happened.  People are also very 
poor at guessing how improbable events really are - again, just as a 
general statement.  Your story, if taken at face value, strikes me as 
improbable too - but the fact that it strikes me as improbable might 
simply not be significant.  Or it might be.  But my intuitive evaluation 
of the improbability is not to be trusted, even as to rough order of 
magnitude.

What permanently zaps the paranormal explanation is not studying physics, 
or studying the history of science, or reading the accounts of debunked 
psychics - it's studying the cognitive science of human error.  The 
literature on this has to be seen to be believed.  The human mind is so 
wildly fragile, so wildly wrong on so many simple problems, that our 
physics is simply more reliable than any anecdote that can be cited 
against it.  *Any* anecdote.  Yes, even anecdotes that really, really seem 
like they can't be explained away.  The human mind is genuinely that weak; 
physics is genuinely that strong.

It doesn't matter whether the explanation I gave is correct.  There's no 
ESP in the universe, no paranormal phenomena, no gods, no demons.  It's 
just us within the laws of physics.  That's the upshot, whatever the 
explanation is - whether you made up the anecdote as a test, or whether 
human memory is horrifyingly (and replicably) pliable, or whether some 
other ordinary event happened.

If you haven't reached that point of confidence, then by no means should 
you attempt to convince yourself of it artificially.  It took me a long 
time to reach the point where I was ready to say that, and it is not at 
all what I was thinking when I started out.  But you know... it really is 
normal.  It's all normal.  And if you look at it long enough, there's 
enough evidence to see that it's normal.  I hope that helps.

-- 
Eliezer S. Yudkowsky                          http://singinst.org/
Research Fellow, Singularity Institute for Artificial Intelligence