How are scientific theories produced

[I_Pwn_Crackpots]

Maybe i have a distorted opinion about science...However, it is possible that you have a distorted opinion and it doesn't matter that you can support each other... The opinion of the majority wasn't always the correct one... However, if your opinion provides results (i have my doubts, but OK), you should insist on it. But I will mention this in the most friendly way:

 

The definitions of hypothesis and theory in a scientific context have very clear, disambiguous meanings. It is not really subject to debate.

 

dogmatism never led beyond a dead end...

The basic problem of the theoretic sector is the confusion about the terminology that is used. You identify theories with verified hypotheses and you think that your approach is correct? I try very hard to keep a low profile and do not be provokative- that's why I won't enter this forum again. Thank you for your participation and, if I could give you an advice it would be: Before adopting the opinions others have expressed before you, take your time and think carefully about all the potentials- all the possible choices.

 

There is no confusion among the scientifically literate about what constitues a theory. I think what is going on is that you are just simply not understanding that most of science isn't opinion.

 

Aristotle has sustained this thousands of years ago, so I let his text speak for itself:

 

Aristotle was wrong on quite a bit of things, so I have no reason to follow any of his advice on scientific matters.

[Mr Skeptic]

I repeat that the fact that we do not conciously know the method we use (and consequently we call it äccidental"or coincidental") doesn't mean that there is no method... Since a theory is produced, some way (yet un-known) is definitely used in the prodution procedure... The fact that we do not know the way, doesn't mean that a way does not exist!!!

 

The fact that there is no specific method required means that they are free to use any method they please. There is not one method as you are implying, but a multitude of methods, perhaps even more than one method per scientist.

 

the reality remains that many scientists "äre stuck" with theories that in the end are proved wrong, simply because they were approved by the majority of the scientific community.

 

Not really. All theories are correct under the conditions that it has been tested on, or they would have been discarded.

 

in the past everyone believed that earth was flat and at th center of the universe. Today we may laugh at this statements, but few centuries ago this opinion was believed by everyone!

 

That's not true. Not only did the ancient Greeks know that the world was a sphere, but they even measured its radius. And if you look very closely, you will find that the world is indeed flat. However, if you look from farther away, you will be able to measure the curvature. What I am saying is that "the world is flat" is not wrong, but it is a first degree (linear) approximation.

 

As for the sun rotating around the earth, that is as true today as it was a few thousand years ago. It is true from an earth-centered rotating reference frame, which is the reference frame that they used back then, and we still use once in a while.

 

Servet, who claimed that heart was just a blood pump ws burnt to death, because everyone else believed that heart was the center of all the emotions!

 

Oh do tell. Were the people who burnt him following the scientific method? There is a reason that real science is considered to have started a few hundred years ago, you know.

 

Since the method is known, how come we do not produce only correct theories?

 

As I said, we only produce true theories, and whatever is not true we discard. However, people assume that the theory will also hold under conditions it has not previously been tested at, and that is not wise. Whenever a theory is being used under conditions that it was not tested with, it is better to consider it again a hypothesis.

 

Eg when science says "kinetic energy is half mass times velocity squared, for speeds of less than 1,000,000 kilometers per second and error margins of 0.0005 %", then some people translate that as, "kinetic energy is half mass times velocity squared for all speeds and to infinite precision" and then act all surprised when that turns out not to be true.

 

Why are there so many unknown phenomena?

 

Cause scientists gave up on saying "God did it" and started demanding a real explanation that actually makes predictions.

[Quartile]

noether's theorem is very important.

[Harlequinne]

The wheel,was quantifiably efficient.Control of the fluke when Adam saw a fallen tree,and thought to himself I'll use its rolling capacity.

The most 'true' aspect of a problem represents a stepping stone.Whether the right or wrong path is represented.A form of clarity ensues.Development thus.

The comprehension of Newtons theory of Light.(linearity)

Young's double-slit experiment.What is it similar to?What do we lean on from experience?How is the same phenomenon in light,suitably revealing to future generations?

The use of logic is of paramount importance.Fluid comprehension,of a justified nature.

A possible order to your question,provided phenomenon comes first,might be:

Phenomenon,comprehension through logic,quantification so that 'order' is proven via the facts.And these facts can be measures of Genius or they can be for the layperson.Their constituent parts are necessarily true.The truest line is always sought.The falsest is just that.It doesn't make it useless,but it must be recognised as 'not an option'. etc...... Nobody said rocket science was easy..just look at Discovery by NASA. That was my design when I almost killed myself as a child!But it is certainly cost effective.

[granpa]

a hypothesis is something you believe may be true. a theory is something you believe is true.

 

but seriously, the key idea here is 'consistency'. internal consistency and consistency with empirically observed data. that seems to me to be the basis of 'inductive reasoning'. you test an idea by seeing if it 'fits' everything else that you believe.

 

you first start with a set of assumptions. then you test each assumption against everything else you know. then based on the outcome you make a set of corrections . the biggest error people make is that they forget that when you get new data you have to go back and retest each of the assumptions that were made originally that led to the new data. you cont just assume that because the old assumptions led to the new data that the new data must support the old assumptions. in doing so people unthinkingly carry false assumptions to ridiculous lengths.

[iNow]

a hypothesis is something you believe may be true. a theory is something you believe is true.

Neither of your definitions are valid in a scientific context. If you were trying for humor, it didn't translate well.

 

 

http://en.wikipedia.org/wiki/Scientific_hypothesis

http://en.wikipedia.org/wiki/Scientific_theory#Science

[pioneer]

Theories are part perspiration (intellectual work) and part inspiration. The inspiration part occurs through parallel unconscious processing of the same data and same analysis. If you look at the two hemispheres of the brain the left brain is more designed for conscious analytical work. The right brain is more spatial-3-D and creative. The left is better at differentiating. While the right is better at integrating. It comes down to being able to translate this right hemisphere data crunching and presenting it in a left hemisphere way, which is the method of science.

 

Let me give an example of the process. Kubule's discovery of the structure of benzene. He was racking his brain trying to reason this structure, mostly left brain, using his analytical skills as a scientist. His right brain was also working on this same project, but unconsciously. It integrated his thought processing into a image of snakes in a circle. Symbols is the native language of the right side of the brain. He then translated that 3-D symbol using left brain language. When this was all done, the published work looks like it was all left brain.

 

Here is a way to visualize the process. The spatial right side of the brain stores data like a 3-D ball. The analytical or 2-D left side of the brain uses 2-D logic planes of cause and affect. To approximate the 3-D ball, we can use a large number of 2-D logic planes, all with a common center. This center is analogous to " what does benzene look like?". The logic planes are the many logical approaches that were used, all with this same center or goal. The 3-D side of the brain makes a ball out of these. That is the snakes in the circle. This is translated into a new logic plane so it can be presented in a science journal, which has a logical and experimental protocol. It may not allow snakes in a circle, even if this made it possible.

 

One way to look at the actual creative process is say we start with four 2-D logic planes being used to make the 3-D ball in the right brain (4 possible ways to explain that lead to nowhere). Relative to the 3-D ball, if we ignore the axis of all the 2-D planes, and look only at the details, it looks like the data is scattered in 3-D. Using 3-D, the brain can appear to jump between planes, illogically, since it got rid of all the coordinate systems but kept the data. One has the answer, "benzene has a ring structure, without any appearance of logic; magic. Ones then needs to translate this back to 2-D or one will never be able to publish.

 

We have a new center, "benzene is a ring" that needs a new plane, since none of the others worked. But the data in all the original planes are 3-D relative to this new plane. But each data will cast a projection or shadow onto the new plane. If you use all the data, it is a more complex line of reasoning that can get esoteric, since the full projection brings down data that does not appear logically consistent in 2-D, but is consistent in 3-D. Interestingly, random and chaos is a way to include some of this 3-D data onto a 2-D plane, without leaving the left brain.

[Sayonara]

Pioneer, you have written five paragraphs without even touching on the OP's question of how scientific theories are produced and advanced. What you have written may make sense to you, but in the context of this thread it is plainly gibberish. Carry on derailing this thread in that fashion and you will receive a spam infraction.

[north]

First of all, I think it's wonderful that people focus their efforts on creating theories regarding various subjects and spend time discussing about them with other people . Nevertheless, I have a qustion to pose: every day we produce several theories during our effort to understand the universe that surrounds

us. Depending on our studies and our interests we elaborate on specific topics, trying to reveal some of the mysteries that "provoke"our mind. But does anyone know (and can tell the rest of us) HOW are theories produced? Many philosophers have tried to answer this question, but the answers that they have given have left me unsatisfied. Production of theories has been attributed to immagination, logics, accumulation of evidence, etc., but no one has EVER provided the scientific community with an accurate, precise mechanism. How can we evaluate whether a theory is right or wrong,complete or incomplete, if we don't even know how it was produced? How can we expect science as a whole to create well- founded opinions (theories) if there is no established mechanism? In order to produce any kind of product (theories are one of the intellectual- mental product) a well established procedure is essential. Otherwise, when the procedure is not known, the theoretical activity is empiric and NOT scientific. I agree that on one hand we do know how to produce theories ,given the infinite number of theories that have been produced during the history of science. On the other hand, how can we explain the fact that no one has expressed at least a rudimentary method? Is it enough for scientists to produce theories without realizing how they manage to do it or should focus our efforts on trying to solve this problem? I simply remind you that many theories , some of which were considered equivalent to scientific knowledge, have been proved wrong or inadequate, despite the fact that they were accepted by the entire scientific community. Is there any chance that such loss of time would have been avoided if we knew how theories are produced and evaluated?

 

yes

 

theories are produced it seems by logic alone

 

sort of the spock syndome

 

perhaps the use of Reason would be a better approach

 

the difference between logic and reason is this ;

 

logic uses the info provided and comes to a conclusion

 

whereas reason accumlates , draws in info all the time , and questions and therefore , at times comes to a different conclusion

[iNow]

yes

 

theories are produced it seems by logic alone

 

sort of the spock syndome

 

This was already addressed above many times, but I suppose it needs repeating. The context of this discussion is "scientific theory," so your comments are far from accurate or true. If you would like to learn more about what I mean, you only have to look back 3 posts to my previous response.

 

Nobody here is talking about scribblings on the back of a napkin in a bar at 3AM after too many beers.

[north]

This was already addressed above many times, but I suppose it needs repeating. The context of this discussion is "scientific theory," so your comments are far from accurate or true. If you would like to learn more about what I mean, you only have to look back 3 posts to my previous response.

 

Nobody here is talking about scribblings on the back of a napkin in a bar at 3AM after too many beers.

 

I see your point

 

but at the same time , this space-time perception , as in both have some sort of reality keeps popping up

 

so that both space and time influences things is pervasive

 

reason says otherwise

[iNow]

Reason also says that chili pepper, salt, and chocolate shouldn't taste good together, yet they're outstanding. Empiricism crushes reason every time.

 

There are a lot of things in our universe which don't make intuitive sense, but you have to remember that our intuition evolved during a very short lifetime as a very small animal on a very small planet in a very small solar system in just one of billions of galaxies in a supercluster of galaxies which is itself one among billions.

 

Trust the experiment, not the logic. As smart as Aristotle was, we've come a very long way since then, and they call it "The Enlightenment" for very good reason. <pun intended>

[north]

Reason also says that chili pepper, salt, and chocolate shouldn't taste good together, yet they're outstanding. Empiricism crushes reason every time.

 

There are a lot of things in our universe which don't make intuitive sense, but you have to remember that our intuition evolved during a very short lifetime as a very small animal on a very small planet in a very small solar system in just one of billions of galaxies in a supercluster of galaxies which is itself one among billions.

 

Trust the experiment, not the logic. As smart as Aristotle was, we've come a very long way since then, and they call it "The Enlightenment" for very good reason. <pun intended>

 

we are now beyond Aristotle and Einstein

 

for a very good reason

[fredrik]

To add another opinion.

(I apologize if I repeat any points already made.)

 

To Martin:

There is no mechanical method for doing science

 

Mechanical or not, there should be a method! Otherwise science couldn't produce so many theories!!

 

The quest for such method and also the difficulties of finding such a method, is pretty much at heart of philosophy of science, in particular the "problem of induction", which questions wether there is a valid method of induction.

 

This largely overlaps with this recent discussion on the philosophy of science: http://www.scienceforums.net/forum/showthread.php?t=36917

 

About "empiricism vs reason" I expressed my personal opinion of that, although dressed in different words in the above thread.

 

A common definition of Empiric is from http://www.thefreedictionary.com/empirics

 

"One who is guided by practical experience rather than precepts or theory"

 

IMO, as is implicit in my take on this there really is no contradiction between the two views. A simple illustration of this is that, for sure a rational reasoning must take into account all available information, including all empirical one! But then empirical data always changes, each time we are consuming new observations. The natural synthesis here is IMO to conclude that there seems to exists no "fixed background reason". Or I think you can also say there is "no mechanical method".

 

But then to respond to

Mechanical or not, there should be a method!

 

It seems that the self-reference here, suggests that since any rational method interacts with itselfs, and thus "evolves". At best one can imagine that there is a rational method for dynamically update the method, so as to stay consistent with the latest empirical data. But one soon starts to suspect that that is subject to the same critics. So we are lead to repeat the argument, and are lead to an expansion, method of method of method.

 

But, at some point, as I argued in the above thread, the reasoner, having limited resources and representation capacity, can not support such infinite constructs. Therefor at some point, it's simply an irreducible opinion, take it or leave it. At this levels, we have an irreducible irrationality. But fortunately this irrationality is tamed by the evolved method.

 

So in conclusion, I think there is a sort of method, but it's self-modifying method, which at some level contains an al least "momentarily" irreducible element of irrationality.

 

/Fredrik

[D H]

The natural synthesis here is IMO to conclude that there seems to exists no "fixed background reason". Or I think you can also say there is "no mechanical method".

This is exactly why I brought up the no free lunch theorems in that other thread on philosophy of science. To say that there is a mechanical method for generating theorems implies, in my mind, that there is such a thing as a free lunch.

[thedarkshade]

In general we look for a new law in the following process: First we guess it. Then we compute the consequences of the guess and we see what it would imply and then we compare the computational results to nature' date=' or we say compare to experiment or experience. Compare it directly with observation to see if it works.

If it disagrees with experiment, its WRONG! In that simple statement is the key to science. It doesn't make a difference how beautiful you get it, it doesn't make a difference how smart you are who made the guess or what his name is. If it disagress with experiment, WRONG! That's all there is to it.[/quote'] That's all there is to it!

[fredrik]

DH, I think we to large part agree. The interesting part, which I consider was part of the "point" in the other thread, which was mainly the goal of my argumentation, is that questions in philosophy of science often lead to interesting and constructive progress for science too. My motivation for bothering, is that often, the opposite view is put forward. So I think it's my responsibility to at least add my opinion to maintain balance.

 

Perhaps by development of method, but sometimes method and subject are similar, in particular is this so for various forms of artificial intelligence and machine learning. And indeed if we're talking about machine learning the problme of induction and rational reasoning is close. My opinion is that there exists no such universal fixed computer on which such algorithm can live.

 

I consider these things to be philosophy of science, but it overlaps with the science of machine learning, because there are simialrities. This relation between software and hardware, is quite similar IMHO at least with the relation between physical law and physical observers.

 

This does relate to the problem of what's to be considered as "observable". Clearly, without observers, the entire notion of observable is ridicilous. YET, we expect the laws of physics to be observer-invariant.

 

/Fredrik

 

The description of this dual view as a bit paradoxal, also reveals what I persoanlly think is the implication of this reasoning, on the view of symmetry principles in physics. IMHO, they are best seen as rules of construction and reasoning, rather than fundamental structural features of nature.

 

/Fredrik

[D H]

And indeed if we're talking about machine learning the problme of induction and rational reasoning is close. My opinion is that there exists no such universal fixed computer on which such algorithm can live.

You are not alone in that opinion. That is essentially what the no free lunch theorems say. To say that the theory generation process is mechanizable is to say that there is a free lunch in the scientific method. I disagree. So, apparently does Einstein ("I never came upon any of my discoveries through the process of rational thinking.").

[fredrik]

You are not alone in that opinion. That is essentially what the no free lunch theorems say. To say that the theory generation process is mechanizable is to say that there is a free lunch in the scientific method. I disagree. So, apparently does Einstein ("I never came upon any of my discoveries through the process of rational thinking.").

 

Now I think I maybe see what you mean and why we keep argue. It seems your main efforts is to argue towards the point that.

 

Q. Is theory generation process mechanizable?

A. No.

 

(This is essentially a critic of the validity of induction)

 

Simply put, I agree on that. But to me, that is not wher the quest ends.

 

In despite of this I find it undeniable that alot of the time there does indeed exists at least quasi-rational reasoning that is impressively efficient.

 

My choice of focus is then how to merge the two "observervations".

 

Q1. Is theory generation process mechanizable? A1. No.

 

and

 

The very often, unreasonable power of various inductive reasoning.

 

Here I think that rationality is emergent. The perfect rationality is unattainable, but the lack of perfectly consistent rationality doesn't imply total irrationality either.

 

I similarly belieive in emergent observable symmetries. Perfect symmetry is also free lunch IMHO.

 

So I think while there is no perfectly mechanical theory generation, there exists some evolutionary self-modifying theory generation. And this makes it plausible the apparent rationality in nature, albeit imperfect, in despite of the lack of a priori reason to expect one.

 

For me the difference this makes is how you view the quest for TOE:s. You may easily say that there will never be a TOE. I totally agree. But that doesn't answer to our quest of increasing our understanding of the world.

 

So how do we view the extensive use of symmetry arguments in theory building in physics? Is it because we belive there is a perfect symmetry, or is it really just a sort of least irrational reasoning? Ie. given thta we accepted tht there is no perfectly rational reasoning, do we look for the as far as we can see, the least irrational reasoning? That's more like I think of it.

 

So I think of reasoning as rational, but with the contraint that there is an irreducible uncertainty in the meausre of rationality. And rationality is also not universal. Two observers may disagree upon what's rational.

 

So I think there is more to this than

 

"Is theory generation process mechanizable? No. "

 

although I agree on that point.

 

/Fredrik

[granpa]

isnt the process of induction identical to error correction (computer science)?

what conditions have to be met to make it possible to transmit a message over a noisy channel without any loss of data?

 

those who arent familiar with error correction should look up 'multidimensional parity". thats the easiest one to understand.

 

what I said earlier about questioning your assumptions (that led to your new data) whenever you get new data comes from error correction.

[pioneer]

For right or wrong I can generate theory continuously. So I have been able to observe how it occurs. The theory often comes before the solution. You go back, and try to fill in the solution with logic and/or data. You then test the theory, which I don't usually do because it is not available. If all works out, you write it up logically and present the data. What everyone else sees is the final affect, which looks like it was done in a rational or empirical way based on the report. That is sort of a special affect. If the scientist said it came to me while I had a beer, and I need funding, he would not get it. If he said, I did this rationally, the system is happier with that. You got to tell the system what it wants to hear, which may not be reality.

 

If Kubule had approached the science community, right after his dream, and said, I had this dream about snakes in a circle, this is the structure of benzene in a nut shell. How would science react to this? He would lose credibility even if he was right. He had to go back and follow the protocol and make it look logical for the correct special affect. I would guess Einstein already knew the answer before the solution. He then had to work on it until he had the smoking gun. Then everyone else marvels at the logic and he keeps his mouth shut, just in case reality creates a defensive reaction. It is the practical reality of the creator.

[D H]

isnt the process of induction identical to error correction (computer science)?

No. Induction underpins all of science. Scientific theories cannot be proven true. The mass of evidence in support of some theory is not proof of correctness in a logical sense; to think that it does is to commit the logical fallacy of affirming the consequent, "If P then Q; Q; therefore P". What experimental evidence can do is annihilate a scientific theory; "If P, then Q; ¬Q; therefore ¬P" is the logical valid concept of denying the consequent, or modus tollens. One lousy piece of contradictory evidence can throw a hypothesis (or even a long-standing scientific theory) into the dustbin of falsified conjectures. Observing thousands of white swans, but never seeing a black swan, does not logically prove the hypothesis "all swans are white".

 

Popper, IMHO, is simply wrong regarding his objection to induction. He ignores the importance of deduction in the scientific process. From the onset, an ad-hoc hypothesis has significantly less merit than does a hypothesis with a solid logical underpinning. Scientific theories are "proven" in a manner similar to convictions in law. A legal conviction requires a logical rationale in the form of motive and opportunity, lack of exculpatory evidence, and sufficient confirming evidence. Similarly, a scientific theory requires a solid logical underpinning, no disconfirming evidence, and sufficient confirming evidence. Popper is right in the concept of falsification. Because of the problem of affirming the consequent, legal convictions can be overturned and scientific theories can be discarded in light of new exculpatory or disconfirming evidence.

 

 

Q. Is theory generation process mechanizable?

A. No.

 

(This is essentially a critic of the validity of induction)

It is a critique of the concept of naive induction, which is in itself a logical fallacy. Popper posed naive induction as a straw man of the scientific method. Scientists don't just make observations willy-nilly. Their observations are guided by extant or proposed scientific theories. Scientific theories, particularly the paradigm-shifting theories, typically do not arise as a an ad-hoc explanation of disparate observations.

 

An example: Noether's first theorem (http://arxiv.org/abs/physics/0503066). The conservation laws and Lagrangian formulation of physics were well-known and well-confirmed at the time Emmy Noether proposed her theory. She didn't need any new observations to tie the conservation laws to symmetry. Through some creative leap she tied the work of Lie, Hamilton, and Lagrange to formulate a new and extremely powerful theorem (not theory).

 

The process of forming a new scientific theory is inherently creative. It is, IMHO, something outside the realm of lower-level logic and Turing machines.

[north]

the beauty of Reason is that it not only brings ALL current information into the fold on any subject but also allows for future info to be applied and thought upon

 

Reason evolves in and of its self

 

logic does not evolve in and of it self

 

because logic is based on the info given

[Sayonara]

Reason evolves in and of its self

 

logic does not evolve in and of it self

 

because logic is based on the info given

 

No, it's because logic doesn't need to "evolve in and of itself". What you propose is a false comparison.

[north]

Originally Posted by north

Reason evolves in and of its self

 

logic does not evolve in and of it self

 

because logic is based on the info given

 

No, it's because logic doesn't need to "evolve in and of itself".

 

explain

 

 

What you propose is a false comparison.

 

how so ?