Scientific Experiments

Grimble

I thought this was the appropriate place for what is essentially (at least to my mind) a pholosphical idea.

Has there been a fundamental change in the way scientific experiments are evaluated?

Has the standard of proof shifted, as indeed it probably has to, due to the phenomena we are investigating?

I was taught at school that the scientific process was to investigate, to evaluate, to form a theory and to test that theory. In particular that one must conclude that every part of a theory happened as envisaged and could have no other possible cause.

But as the scope of scientific theory continues to grow we focus on the extremes: the 'infinitely' small, sub-atomic particles; the 'infinitely' large, cosmological entities and processes; the 'infinitely' brief, etc, etc.
We seem to be tending towards the line of argument; if A then B and if evidence is found, that can be cosidered to comply with A, then B is considered to be proven.

Please understand that I am not in any way decrying modern science, for perhaps this has always been the way?

The example of Phlogiston comes to mind: first given that name by Georg Ernst Stahl, in 1703 and said by some to have had negative weight, the theory lasted for nearly 100 years.

Grimble

kote

Good question! You're definitely talking about philosophy here, or more specifically, philosophy of science. Philosophy of science asks questions like "what does science tell us?" and "what are the limits of science?"

Many people have spent their lives trying to answer these questions, and we still don't have consensus. The short answer is, however, is that you have been lied to . You'll find that this happens a lot actually, so it's good to ask what's really going on. The fact is that we don't really know as much as we say we do about a lot of topics, but in order to teach anything without getting bogged down we just have to pretend that some things are true when they are deemed close enough.

Science works on the principle of induction. We assume that if we take a measurement repeatedly we will always get the same reading. We assume that the same principles of nature that cause us to measure Earth's gravity to be 9.8m/s^2 at some locations will hold at all locations. We then further generalize and predict what the acceleration due to gravity will be on planets we have never been to. The problem with induction is that we can never have absolute proof that things will always hold up how we think they will. This is known as the problem of induction.

Furthermore, we can always come up with multiple alternative theories that fit experiments equally well. There are always other possible choices. Are you familiar with Galileo and the copernican vs ptolemaic models of the solar system? Both the heliocentric and geocentric models fit experiments equally well. We ended up switching to thinking of the sun as the center of the solar system, rather than calling earth the center of the universe, purely because the math is more simple when we put the sun at the center.

In philosophy there is a famous thought experiment about "grue." Grue is a property of objects which causes them to be green before some year t and blue afterward. Many of the objects you had theorized were green might actually instead be grue. You have equal experimental evidence for both the theories of green and grue. From Wikipedia:You ask another question about having only indirect experimental evidence to support our theories. This too has always been the case, although is is more clearly evident in modern science. When you use an electron telescope to "see" something, you are relying on all of our theories involving how that technology will work. When you just step on a scale you are relying on our theories of gravity to present you with a proper measurement of mass. The form of argument you gave though, if A then B and A is true so B is true, is perfectly valid. If, in fact, you have proven A, then you have proven B without a doubt. The problem is, as discussed above, you can't ever deductively prove any scientific theories (although you can prove them wrong).

apeiron

This is a very interesting question. And we could talk about both the theory and the practice.

The theory of how to do science is still pretty much what it always was. Though the limits are better appreciated perhaps.

Science is modelling. It is a mistake to think it is purely a search for "the truth". It is about creating general models which generally predict the world - or more accurately, predicts the kinds of things we are most interested in predicting. So modelling is tied to human purposes. And the main purpose (that gets socially rewarded) is the making of machines. Technology. Control over the world.

This would then be one kind of change perhaps - that theory building has become even more closely tied to the technological fruits that may result.

There is also the status research, the national prestige research - men on the moon, supercolliders, etc. This is science tied to political purposes - though there is good spin-off for military and economic purposes when it comes to nuclear physics, space races, etc.

Well, already I'm talking about the practical changes. But then there is a whole other kind of change in the fact that there are vastly more academics doing "research" at universities. And that so much of the straightforward science has been done. You know have a situation where multiple standards of quality apply. So there are domains of high rigour and others where anything goes. It is just a much more varied eco-scape these days. It is hard to speak of "science" as a monolithic or homogenic discipline and judge it good or bad.

Returning to your key point, which seems to be about the difficulty of making the measurements that confirm the theories, yes, at the extremes we are getting towards the situation where we can essentially make only the one measurement. The state of the universe we are within.

And people also dream they may discover a ToE - which I see as a measurement-less theory of reality. Just from considerations of pure symmetry, there may be reasons why we live in this 3D realm with certain particle symmetries, etc. A few constants might be randomly chosen as the way this symmetry breaks.

Would this still be science? Well it would probably not deliver us much in the way of new technology. It would probably not be testable by new kinds of measurement (though it would have to fit all the existing ones).

But it could be more like science as the pursuit of pure truth. Indeed, more like philosophy!

And personally I don't see that as a bad thing. I would spare a few taxpayer dollars to fund it.

And with the wrangles over the status of string theory and anthropic principles, we can see people trying to thrash out the rules for entering this next phase perhaps.

JoeDawg

This actually touches on a pretty major disagreement in the philosophy of science. Which comes first, evidence or hypothesis? Its a problem with formulating the method.

This is related to the rationalism vs empiricism argument, as it deals with what inspires a theory. Some theories do indeed seem to come from our imagination... and then get proved or disproved via evidence. But at the other end of things, sometimes scientists are inspired by the evidence, that is, the evidence points to the theory that no one thought of before.

Its really a chicken and egg problem. Scientists are always evaluating evidence, but they can also be inspired by seemingly unrelated things.

The use of 'double blind' experiment also comes to mind as a 'development' in how science is done.

kote

Oh, so is psychology a science now ? I think that's the development.

apeiron

Ah, well at least that is always an easy loop to get out of.

Understanding begins in a vague state of development and becomes crisply developed - dichotomised into generals and particulars, models and measurements. The two naturally go hand in hand as in synergistic activity, each driving the other in their opposed directions. So no need to worry about which comes first.

JoeDawg

Both are, but psychology is just a new field of study, its not a change in method. You can apply 'scientific method' to anything really, including economics, and even astrology.

Double blind is a good example of how to control for bias, and the importance of the impact of the observer, has been something that has come slowly to science. Even now, physicists are grappling with this in terms of quantum theory.

JoeDawg

That doesn't really address the problem of defining method.

apeiron

This is the method....

Understanding begins in a vague state of development and becomes crisply developed - dichotomised into generals and particulars, models and measurements. The two naturally go hand in hand as in synergistic activity, each driving the other in their opposed directions.

magpies

No there has not been a fundamental change in the way we do science. While yes I do agree one is needed.

Grimble

Thank you Gentlemen, (No, I must not assume that, must I?), Gentlemen: Good People, one and all, it is a much broader subject than I had anticipated, and deeper too.

In my original post I expressed the following opinion: To which I received the following reply: (thank you, kote, and thank you too for a good summary of the difficulties, your reference to the fascinating problem of induction was very welcome.

But I had mixed up my own line of reasoning there; what I should have said was: We seem to be tending towards the line of argument; if A then B and if evidence is found, that can be considered to comply with B, then A is considered to be proven.

An example of this would be: 'if a star has planets, and if their orbit is in our plane, then that star's brilliance will diminish regularly. Then if such regular dimming of the star occurs, the star has planets'.
(please understand that I am not saying that this is what scientists have said, but it is the way it has been reported) and my question would be "but is there anything else that could give rise to the same evidence/observation?"

Grimble

WaveJumper

Interpretations of what a physical finding means, will always be contested. If A then B, and if B then A will hold for as long as we find a contradiction. When we find one, our assumptions on which a scientific investigation is always based, will have to be revised. You can say that this is what is currently going in physics.

JoeDawg

That is your opinion. There is a wide array of opinions within the philosophy of science on this particular topic, and there has been a distinct shift in opinion over the last several centuries.

And really, you haven't said much of anything about an actual method. If anything, what you have said indicates a lack of discernable method.

One of the problems with defining any method is that scientists often find the greatest success by being unorthodox... while others work with the tools that function best in their field, which means there are either many different methods, or none at all, that is, anything that generates success... wins.

WaveJumper

I agree with this. I have to look up who said something to the effect of:

The greatest barrier to new knowledge in science is its past success.

disregardthat

Apeiron, what do you mean by "crispness"?

apeiron

I use it specifically as the complementary term to vagueness, so it is a technical term.

It would mean definitely and clearly existing, or fully developed, present in most certain form.

The point is that most people assume that all reality IS crisp. Something either exists or it doesn't. So crisp would be a redundant term because that is simply the way things are. Any vagueness would be semantic - as in the sorites paradox.

But I am interested in logics founded on vagueness - the dichotomous separation of pure potential. So the logic of Anaximander, some versions of Tao, and of CS Peirce. Vague and crisp are two ordinary english terms that seem to come closest to capture the essence of the technical ideas involved.

I picked them both up from Stan Salthe who uses them in hierarchy theory approach.

apeiron

It is an opinion supported by argument and shared by others. I've cited Rosen's modelling relations and anticipatory systems work here often enough.

Your problem is that you have not once either shown the basic idea of dichotomisation is in error, or provided some arguments and references in support of some other view.

I can only conclude you don't actually study current epistemology and are only responding from a casual layman's understanding of the issues.

So here are three refs for you to chew on.

The first compares Popper and Rosen. The second is a general intro to Rosen. The third is about bottom-up~top-down, vague to crisp, dynamic logic, based on Grossberg's anticipatory neural nets.

http://www.osti.gov/bridge/purl.cove...u/webviewable/

http://www.people.vcu.edu/~mikuleck/PPRISS3.html

http://www.scitopics.com/Dynamic_logic.html