[ExI] Judging radical possibilities

[Anders Sandberg]

Tomaz Kristan wrote:
> > Einstein was wrong don't notice that there is an even more plausible 
> conclusion if these findings are real: causality is wrong.
>
> It's far more likely that Einstein was wrong than the causality, IMO.

This is an interesting statement. Let's (for this thread) drop the 
question about what we actually think of FTL or relativity, and instead 
look at the quite interesting meta-question "How do we judge the 
likeliehood of radical changes of physics?"

Assuming we were to have to either drop relativity or causality, but did 
not know which, what method would we use to determine the most rational 
course of action? Obviously we would try acquire relevant information to 
make the choice, but if that was not forthcoming it seems we would use 
other principles.

One would be simplicity: which of the two is the most complex or 
introduces more concepts? But both can be formulated in very succinct 
ways with simple postulates.

The common sense approach, basing judgements on past experience, might 
seem to give an overwhelming support to causality rather than 
relativity. But this is likely suffering from serious parochialism, 
since we live in an environment where relativistic effects are minor. On 
the other hand, it is not clear what environment would correspond to a 
"neutral" view of the laws of physics.

Looking at a review of the topic like
http://plato.stanford.edu/entries/causation-process/ and
http://plato.stanford.edu/entries/causation-backwards/
it might look like causation is on shakier ground because there are so 
many different schools of thought, and arguments against retrocausation 
mainly seem to build on avoiding paradoxes. But having multiple 
interpretations or models just shows that it is a pretty active area; 
the lack of consensus might be due to lack of data to force a particular 
conclusion. In physics this is usually bad news: there is floppiness in 
the theory, and Popper et al. will claim it is unfalsifiable or at least 
a worse theory than one with less freedom. But in philosophy this might 
be a less strong argument: we might just not have found the right way of 
expressing ourselves that unifies the different takes on the topic.

The proof by paradox is in worse shape. Just because the consequences 
are nonintuitive doesn't mean they cannot happen. Just consider much of 
modern physics. If we are talking about radical new possibilities we 
should expect unprecendented things that are outside our normal 
experience. However, most people would think that true paradoxes cannot 
happen: we can't get actual contradictions in the laws of physics, and 
hence theories suggesting them must be wrong. But we accept some 
contradictions in our understanding because we think the real state of 
the laws of nature is contradiction-free and our theories are just 
partial models (standard example relativity and quantum mechanics). We 
also often find that other laws of physics intervene to prevent 
contradictions: Novikov's self consistency principle seems to show that 
quantum mechanics prevents many time-travel paradoxes by giving them 
zero probability. The reductios do not take this into account, and hence 
fail. While reductio ad absurdum works fine in math (where intervening 
extra laws cannot happen if we make a good proof)  it might not work 
well on physics at all.

In the end it seems that knowing one possibility is wrong and the other 
right (or, with some small probability, both are wrong and the world is 
even stranger) leads to a situation a bit like moral uncertainty: you 
think there are moral rules you should follow, but you are not entirely 
certain which are the right ones. As long as you keep to situatons where 
both give the same results you can act with confidence. But elsewhere, 
finding good rational action strategies is tough. Some of the ideas in 
moral uncertainty theory might be applicable, but it seems that the one 
thing all theories support is the gathering of more information. Which 
might be easier in physics than in ethics.

(however, if we actually had reason to think we could get metaethically 
relevant information from some experiment, it would probably be more 
important than any physics experiments we could imagine, since the 
information would promote acts of true value directly. The LHC will just 
tell us a bit about how the world is, while a Large Ethics Collider 
would tell us a bit about what we ought to be doing (if anything). )

-- 
Anders Sandberg,
Future of Humanity Institute 
James Martin 21st Century School 
Philosophy Faculty 
Oxford University