mirrorshard: (Lammas print)
[personal profile] mirrorshard
This is something I've had percolating for awhile, sparked off originally by a reader's letter in the Metro, which was (of course) batshit insane. It argued, as far as I can recall, that scientists were right in being more certain and dogmatic than religious leaders. (It isn't a matter of right or wrong; it's down to understanding where certainty comes from and why being certain about anything complicated makes you look like an idiot.)

Since then, I've had a few conversations which have brought new viewpoints to it, most of which ended up with an agreement that the axiom mismatch meant we weren't getting anywhere.

Of course, acknowledging that an axiom mismatch exists is the first important thing. One point which is commonly overlooked is that any system of thought is an artificial structure, erected within the extremely complicated perceptual universe that exists (FSVO, &c.) around us. It depends on context and axioms - the fundamental truths you need to take for granted in order for your system of thought to work. (Ref. Godel.) If you're going to shore up your belief system (your tools for understanding the universe) with hard-edged logic, then Godel's theorem applies; the mathematics are inexorable.

This structure (or scaffold) can be viewed two ways. The first is that it's a good and accurate tool, across the entire domain where the axioms are true. The classical example here is Newton's Laws of Motion. They are completely and exactly accurate for macroscopic objects; they break down and do not apply at very small scales. The problem is that your axioms may fail unexpectedly; the corollary of that is that any system of thought is i) only as strong as its weakest axiom, and ii) strong in inverse proportion to the number of axioms it requires.

The other way to look at it is as a model. (The money quote from there is from Eykhoff: a representation of the essential aspects of an existing system (or a system to be constructed) which presents knowledge of that system in usable form. The map is not the territory; the map is a simplified version of the territory that we can understand and manipulate.) Modelling is a really useful way to deal with the universe, but two things must be remembered: i) no model is complete, and ii) models fail by not taking things into account. So when a model attempts to predict things that are affected by a factor it doesn't consider, the results can become wildly divergent from reality just a short distance in phase space from places where they are effectively exact.

One interesting corollary of the model viewpoint is that you can tell how useful a given model is (ie. how close it is to reality) by looking at its complexity. A complex model of a complex system may not necessarily be accurate, since it might be built wrongly; but a simple model of a complex system will always be wrong.

If you believe that either your axioms are always and forever true, or that your model is complete, then you are no longer using any form of logic known to science and I don't really have anything to say on the subject; your belief is not disprovable. It's not even wrong.

This is one reason that scientists can come across as overly certain and dogmatic; it's because, given the boundaries of the domain they're working in, there is no other possible answer. It's like saying that 2 + 2 = 4; the way the question is phrased requires that there is only one answer. For scientists, the axioms and assumptions are implicit. If you find the arrogance offputting, try mentally appending "assuming the Earth doesn't turn into a very large haddock" to anything that looks like a flat statement of dogmatic certainty. Generally, a competent scientist or engineer will be very careful about their assumptions: 2 + 2 = 4 given nonweird mathematics, and "this bridge will not fall down within 25 years under normal load conditions".

There is, of course, a difference between those two statements; the first is interesting in some contexts, but is otherwise intellectual wankery. The second, if incorrect, kills people. If you bring out your assumptions in the open, and invite people to test them (to try and disprove them) then you reduce the chances people will be killed.

Religious statements about the nature of the universe and our relationship to it are sort of analogous, except that they generally try to hide the axioms & assumptions, and often get terribly upset when those are brought into question.

This is one reason I try and avoid making statements on religious grounds, or trying to affirm something religious; I don't have the degree of certainty necessary to make precise statements. It's like the magic image-enhancement software they use on CSI. I know that some actions have more beneficial results than others, and that some activities give me the peaceful, joyous feeling others have described as numinous or (in a Christian context) as indicating the presence of the Holy Spirit. I know that there is a common characteristic of good things and actions, which I'm comfortable terming "God".

Many religions (and non-religious belief systems) have advocated these actions, and described that feeling; I don't know enough to say that one is more accurate or somehow more correct than another. I consider myself Christian because it gives me a comfortable set of metaphors and a cultural vocabulary to talk about these things, and a Quaker partly for familial reasons and partly because the Society of Friends emphasises aspects of Christianity I find important and rejects some aspects I dislike. (I'll happily talk about those at more length another time, if anyone's interested.) I don't in any sense find truth in religion; I find ways to examine and talk about truth.

One of my sidebar quotations is from Niels Bohr - Never express yourself more clearly than you think. This shouldn't be taken as implying that if you don't have crystalline clarity and precision of thought, you are somehow lacking; the universe is not a clear or precise place. Every time we learn something new about it, and about ourselves, it gets less so. We have methods of dealing with chance and human foibles, but mostly things come down to modelling - to reducing the complexity of the physical, psychological, and social universe (at this point you can wave your hands around a bit and throw the word 'fractal' around, if you like) to something we can handle.

Date: 2008-09-16 10:38 am (UTC)
From: [identity profile] mirabehn.livejournal.com
I will probably find some intelligent things to say about this at some point today.

For now I will just say that you are fantastic and that I love you. :-)

the universe is not a clear or precise place. Every time we learn something new about it, and about ourselves, it gets less so.

*could not agree more*

Date: 2008-09-16 10:39 am (UTC)
From: [identity profile] battlekitty.livejournal.com
I may be missing something here - the only difference between the two views of the structure you're mentioning are whether it's a theory/law or an observed relationship (blah - can't brain today. There's a proper word for this, I'm sure...). Aren't these are different parts of the same thing, but with different levels of reliability? Sure, theories are more elegant and harder to misuse than empirical models, but the empirical model is (in part) a stepping stone to developing the more elegant theory.

The other thing that sprang to mind is that I've found that always stating the restrictions to every statement just leads to me sounding like I'm not sure what I'm talking about. Or just takes too long. (I've mentioned I'm terrible in job interviews, right? "Yes, I can pick this up really quickly, as long as a meteor doesn't crash into the earth destroying thew entire structure of our civilisation. It might take me longer then..." *sigh*)

I hate absolute statements. I don't believe in the existance of perfection. That underpins every thought/belief I have on Life, the Universe and Everything.

I also have a headache and am tired, so sorry if I'm not making sense. :)

Date: 2008-09-16 11:58 am (UTC)
From: [identity profile] mirrorshard.livejournal.com
[T]he empirical model is (in part) a stepping stone to developing the more elegant theory.

This is quite right, when you're developing the theories in the first place, because of course nothing helps so much as a good theory - but when applying a set of models and theories you've already got, sometimes it helps to take a step back (or at least sideways) and use the model rather than the theory.

The other thing that sprang to mind is that I've found that always stating the restrictions to every statement just leads to me sounding like I'm not sure what I'm talking about.

This apparently comes under Communication for Scientists (Advanced Course) - otherwise known as talking-to-management. I've never managed to get the hang of it myself properly, but I find that mentally appending an "unless something really odd happens" helps.

I don't believe in the existance of perfection.

Nor do I, at least in any interesting non-theoretical sense. Perfectibility is a bit more tricky, but I like to think of it in terms of improvability instead - that things can asymptotically approach perfection, and there's always room for improvement.

As in most interesting areas of life - reject hard binaries!

Date: 2008-09-16 12:26 pm (UTC)
From: [identity profile] battlekitty.livejournal.com
...I don't believe in the existance of perfection. [+ your response]

Agreed, but definitely stressing the asymptotically...

In passing, that statement is also the basis of why I first stopped believing in the Christian (etc) God - it requires too many absolutes in a world that is anything but. Hell, not even acceleration due to gravity (at sea level, standard temperature and in a vacuum that can be taken as acting as absolute - have I got them all? No weight* wait: On Earth!) is an absolute value...

*facepalm* I was fine until I had to try remember all the conditions. *sob*

A lot of the global warming science has suffered from these sorts of problems, too - one of the founding concepts in the Rio de Janiero Conference (from memory - I'm probably wrong. The conference that set up the UNCCC/put forward the need for the Kyoto Protocol) was that the risk/consequences were much too serious for the governing bodies of the world to be able to wait for proof before acting. (Unfortunately, they have.)

* I really did type that accidentally, but I thought you might appreciate the irony... :)

Date: 2008-09-16 11:19 am (UTC)
From: [identity profile] atreic.livejournal.com
Yay! :-)

[I feel this post requires a proper comment; then again, I feel interesting and honest long posts tend not to get any comments because commenting on them is difficult. So you can just have a yay, this post was interesting and beautiful]

Date: 2008-09-16 11:33 am (UTC)
From: [identity profile] mirrorshard.livejournal.com
Thank you!

Date: 2008-09-16 11:26 am (UTC)
From: [identity profile] friend-of-tofu.livejournal.com
[livejournal.com profile] alextiefling is going to be gutted you've posted this - he's been all geared up to write a big essay about uncertainty...

Top notch post. I take my hat off to you.

Date: 2008-09-16 11:33 am (UTC)
From: [identity profile] mirrorshard.livejournal.com
Thank you! And I'd be really interested to hear Alex's view on it in any case.

Date: 2008-09-16 05:57 pm (UTC)
From: [identity profile] alextiefling.livejournal.com
It's on its way - don't worry. You haven't exactly pre-empted me, nor will I try and steal your thunder by saying things that are too much like what you say here. But there's a definite overlap.

Date: 2008-09-16 07:21 pm (UTC)
From: [identity profile] yvesilena.livejournal.com
[livejournal.com profile] mirabehn was right, brilliant post!

Date: 2008-09-16 07:39 pm (UTC)
From: [identity profile] mirrorshard.livejournal.com
Thank you!

Date: 2008-09-19 12:34 pm (UTC)
From: [identity profile] jamesofengland.livejournal.com
Got here trying to follow a reference from [livejournal.com profile] wildeabandon. Was very impressed.

Date: 2008-09-19 12:58 pm (UTC)
From: [identity profile] mirrorshard.livejournal.com
Thank you - I'm glad this has spread and been interesting.

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags