Sunday, October 28, 2007

Science Have Its Limits....?

Science is a method, it requires no faith. In fact it is a method through which provides it's own falsifiable test of itself.

Slow down there, cowboy. Nothing proves itself -- you always start with a certain set of axioms.

While it is indeed one of the great tools for knowing things that we have, it is certainly not the only way things become known. We can learn certain things through reason alone (such as math), and many things can only be learned through word of mouth (Sally said that Harry said that...). Statistics is one of the fundamental answers to epistemology (how can I know something), but ultimately we only can learn things at certain (not very high) confidence levels. While a p-value of 0.05 or 0.01 might sound pretty impressive (and are the standard rules of thumb for statistical 'proofs'), they represent 1-out-of-20 and 1-out-of-100 studies' results being nothing more than the result of random chance. If you have, say, 10,000 papers published a year, 500 or 100 of them will be wrong.

Given how often scientific answers have indeed been found to be wrong, especially in epidemiological studies (which is a sort of scientific wishful thinking), it hardly proves itself to be true (which can't be done anyway). A better way of putting it is, "It's the best method we have of figuring out empirical truths about nature."

There are very major limits on science and the scientific method. Notably:
1) Singular events. Science can't handle singular events very well, or not at all. For example, suppose the people that claimed they had seen cold fusion back in '89 really did see Cold Fusion. Perhaps a gamma ray hit something at just the right time, or maybe it required high altitude, or something. But when researchers tried to duplicate it, they couldn't and so the guys were branded as frauds. Maybe they were, maybe they weren't... but they could actually have made an honest empirical observation, and then branded as frauds as a result of it.

2) Trust. The motto of the Royal Society is "Nullis in Verba" ("On the words of no one") In other words, don't believe what people say, but only trust in reproducible experiments. The trouble with this is, of course, that no one can come close to reproducing all of the empirical experiments needed for a full understanding of modern science, and so it always boils down to trusting what other people say. If a car full of scientists drove through a mountain pass and saw a white substance outside, they could send one of their members out to report if it was sand or snow... without accomplishing anything. The friend could be playing a practical joke on them, after all. All of them would need to go outside and make an empirical observation of the substance themselves in order to be satisfied. This is a very fundamental flaw in the system, which only works since malicious papers (as far as I know) are not inserted into the literature like viruses.

3) The old induction problem / uncertainty. Science is based on inductive reasoning, and inductive reasoning from empirical events can't actually prove anything. We can make certain claims, but not proofs in the sense that logical or mathematical statements can be proven true. "The sun will rise tomorrow" is a scientific claim, but it cannot be proven to be true. The fundamental problem is that what is true in the past might not be true in the future. Since certain things like universal constants are likely to stay the same (though some have theorized they have not in the past!), it can be answered by simply stipulating "If things stay like they are now..." but this is still not the same level of proof as people deal with in logic and math. All scientific knowledge, ultimately, is uncertain.

4) Heretics. The heretics of science have always received rough treatment. Most of the time it is deserved (there are a lot of nutcases out there), but sometimes people have followed the scientific method but had their papers rejected because the reviewers assume their preconceived conclusion. The guys who discovered that stomach ulcers were caused by heliobacter pylori had their landmark paper rejected because it went against what was commonly known to be true -- the rejection of which was the entire point of scientific knowledge.

However, for all its flaws the scientific method is, as I said, the best method we've found so far of determining empirical truths about nature. It is certainly not the be-all and end-all, as by its very nature it will reject all unique events. It is fundamentally useless for knowledge about anything which cannot be reproduced, and what it does know it only knows with varying degrees of certainty. It cannot say anything about anything outside of its sphere of influence (empirical observations of the natural world), and so stands mute on being anything but a fact-provider to fields like philosophy, religion, logic, math, ethics, and et cetera.

The scientific method certainly doesn't deserve the religion-like attitude of worship Popper and many people on here seem to give it.
Posted by Ajay :: 12:59 PM :: 0 comments

Post a Comment

---------------oOo---------------