Jump to content

How are scientific theories produced


Effie
 Share

Recommended Posts

Shannons theorem relates the maximum capacity given, signal to nosie ration, the noisy channel theorem says

For any ε > 0 and R < C, for large enough N, there exists a code of length N and rate ≥ R and a decoding algorithm, such that the maximal probability of block error is ≤ ε.

And: N -> inf, as ε -> 0.

 

the set of everything that we know about everyting may not be infinite but it is very very large.

 

certainly if every bit in the message has a chance of being wrong then its possible, howevery unlikely, that every bit in the message is wrong. no error correction can repair that. but what if the total number of errors is known to be less than some maximum?

 

and what if R<<C?

Edited by granpa
Link to comment
Share on other sites

Rational theory has to be almost perfect. If someone can find an exception to the rule, either the theory needs to be revised, or it will be replaced by a new theory that can also take the exception into account.

 

Empirical and statistical based theory gives itself more exceptions than one, before it is considered wrong. It uses a fudge factor, that says this theory is allowed to be wrong + or - exceptions. If it was rational theory it would only be given one exception. Therefore it is irrational theory, since more than one exception has little impact compared to rational theory.

 

Here is an example of cutting rational theory as much slack. After careful observation of 100 swans in my lake, since they are all white, I conclude all swans are white. My neighbor scientist sees the lake, my observations and my logic, but my theory does not sit well. He decided to challenge my theory and find a gray swan. In the rational world, I have been disproved. I was king of the hill for a couple of days. But in the statistical world I am still in business.

Link to comment
Share on other sites

how does that address the original question?

 

what if the swan is 99.9999% white? how are you defining 100% white (or white in general)?

 

white is a category. it includes a range of colors. 99.9999% white is white.

(not to mention that inclusion in a category is itself not all or nothing)

 

 

further it is obvious to me that all this talk about swans is a straw man argument. we probably couldnt agree on whether all swans really are white or what the definition of white even is. a much more pertinent question would be how we know the sun will rise tomorrow. we all know it will but how do we know? it obviously isnt a simple matter of observing it happen 1000 times then concluding that it must happen tomorrow.

Edited by granpa
Link to comment
Share on other sites

how does that address the original question?

You need to look in a mirror, granpa.

 

You have made six off-topic posts in this thread: This one, this one, this one, this one,

this one, and this one.

 

Your posts are off-topic as Shannon's theory is a syntactic information theory. This thread is about semantics. Semantic information theory is a different beast than syntactic information theory.

 

Shannon's theory proscribes means for transmitting the contents of a message over a noisy channel. Shannon's theory does not care about meaning of the message. It is quite happy in getting zero transmission errors after successfully sending and receiving the message "A triangle has four sides." This message has null semantic content.

 

Another between Shannon's theory and semantic information theory is the location of the "truth" data. Because syntactic information theory doesn't care about the semantics of a message, the means for detecting and correcting errors must be embedded in the message. The truth is not embedded in a scientific thesis. The thesis must instead be tested against external reality.

 

Shannon was very clear in his development of his information theory that he was not addressing information theory as a whole. From Shannon, C., 1953, "The lattice theory of information", IEEE Transactions on Information Theory, 1(1), 105-107: "The word 'information’ has been given different meanings by various writers in the general field of information theory. It is likely that at least a number of these will prove sufficiently useful in certain applications to deserve further study and permanent recognition. It is hardly to be expected that a single concept of information would satisfactorily account for the numerous possible applications of this general field."

 

Some reading on semantic information theory:

 

Carnap, R., Bar-Hillel Y., 1952, "An Outline of a Theory of Semantic Information", TR 247, Massachussetts Institute of Technology.

https://dspace.mit.edu/retrieve/4903/RLE-TR-247-03150899.pdf

 

Floridi, L., 2004, "Outline of a Theory of Strongly Semantic Information", Minds and Machines, 14(2), 197-222.

http://philsci-archive.pitt.edu/archive/00002537/01/otssi.pdf

 

Floridi, L., 2005, "Is Semantic Information Meaningful Data?", Philosophy and Phenomenological Research, March 2005

http://philsci-archive.pitt.edu/archive/00002536/01/iimd.pdf

 

Also see http://plato.stanford.edu/entries/information-semantic

Link to comment
Share on other sites

Dear granpa,

 

You have received an infraction at Science Forums, The Original.

 

Reason: Flaming

-------

Your surliness and insults do nothing to help make you understood.

-------

 

This infraction is worth 10 point(s) and may result in restricted access until it expires. Serious infractions will never expire.

 

Original Post:

http://www.scienceforums.net/forum/showthread.php?p=455565

 

thats what I figured. you have no idea what I'm even saying.

 

ignorance is bliss. enjoy it.

 

All the best,

Science Forums, The Original

Link to comment
Share on other sites

threatening to excommunicate me is hardly a threat. I could care less.

 

I have the impression that no matter how simple and obvious something I say is that its going to be endlessly picked apart by at least one of you. and if you cant argue with what I said then you just pretend I said something else and argue with that.

 

I can get a little cranky sometimes and I probably shouldnt post when I'm in such a mood. I apologize for that.

 

I have posted many thought provoking ideas on this site hoping that I might find some others that also find these things interesting. but nobody seems to be interested in having any intelligent conversation. everyone is only interested in selling their own personal pet idea and in tearing apart everyone elses ideas. this is whats wrong with, not just this site but, the scientific establishment and indeed the whole world as far as I can tell.

 

if you excommunicate me then so much the better.

Link to comment
Share on other sites

I have the impression that no matter how simple and obvious something I say is that its going to be endlessly picked apart by at least one of you. and if you cant argue with what I said then you just pretend I said something else and argue with that.

 

...

 

I have posted many thought provoking ideas on this site hoping that I might find some others that also find these things interesting. but nobody seems to be interested in having any intelligent conversation. everyone is only interested in selling their own personal pet idea and in tearing apart everyone elses ideas. this is whats wrong with, not just this site but, the scientific establishment and indeed the whole world as far as I can tell.

 

Amusingly, you just described the standard crackpot. They come, wanting to share their crazy ideas with everyone, and not interested in anyone else' s ideas (including and especially the ones accepted by the scientific community and backed up by lots of data). They are not interested in intelligent conversation -- if someone criticizes their idea, they ignore it, or make excuses, but usually try to distract everyone by arguing about something else they said.

 

While crackpots are useful for coming up with new ideas, the ideas are useless until a scientist goes over it and does all that work that the crackpot was unable to or incapable of doing. Unfortunately, crackpots rarely appreciate the scrutiny involved and tend to leave when someone starts asking them for details or evidence, or any criticism they can't answer.

Link to comment
Share on other sites

I have posted many thought provoking ideas on this site hoping that I might find some others that also find these things interesting.

You are encouraged to do so, but you must keep those "provoking ideas" within the realm of science and on-topic. While Shannon's information theory most certainly is within the realm of science, it is not germane to the topic at hand. First and foremost, it has nothing to do with how theses are generated. Secondly, it has nothing to do with how theses are evaluated.

Link to comment
Share on other sites

Looks like someone disagrees with himself. :D

 

Fortunately it's not as bad as it looks :P Sometimes questioning yourself is the way to make progress.

 

In the past posts I was try to be diplomatic. Granpa associated the original topic to "error correction", and while I think that the classical shannon theory doesn't near solve the problems, I see a way to make reasonable associations in a wider perspective. Everyone is coming from a different direction on here and appropriately to the discussion, I can only guess what any poster is trying to communicate :) And to apply statistical inference, I first need to guess the channel, which isn't known either.

 

The original question was how scientific theories are constructed. How is that interpreted? Is it a history question? or is it an attempt to create an abstraction to some kind of "logic of creativity". I think the obvious answer that noone has a eterministic "method of creativity" was given early in the thread, yet the discussion lived on.

 

In my personal abstraction of predictions and how it relates to communication channels is vaugely this

 

Consider and abstract observer (could be a physical system, not necessarily a biological system). What can this observer infere about the reality in it's environment? One can picture that ther is a "comunication channel" to the environment through wich information is fed. And the problem is, given the observations, what can it infere about the outside?

 

Then, based on our "expectations" on reality, and on the "future" a rational observer can determine what action to take.

 

But this abstraction raises many questions that invalidates the abstractions in the original shannon theory.

 

In shannon theory, inference and error correction is possible statistically if the channel (transition probability) and marginal probability is given. In the above example of mine, these are not given, they also needs to be guessed.

 

In my personal abstraction, this relates to learning. Learning, as in related to science (learning about nature, by means of processing experimental input and experience) means, developing the opinon of the "communication channel" through which you interact with reality, in a constructive way. This further limits the possible error correction because there is uncertainty also in the channel (transition probabilites).

 

Anothre problem is when the receiver is saturated. A finite physical systems can (most would agree at least) hold only a finite amount of information. Therefore, the abstraction of asymptotic stead state streams is even more inappropriate. This finite information capacity in the nodes, implies a kind of "cutoff". Which further introduces choices: What information to discard, can we compress the data etc?

 

This is how I see a possible abstraction to learning. So to realte to the error correction of shannon, a learning model, is not just error correction over a knonw channel, it contains feedback whereby the channel evolves. The evolution of the channel is analogous IMHO the evolution of questions asekd, and the evolution of experimental design.

 

In a certain sense an experimental setup, does specify a kind of "communication channel" (although subject to various uncertainties) through which we make inference about nature. But these communication channels are not given! WE built them!

 

/Fredrik

 

x -> [ noisy channel ] -> y, characterized by p(y|x) - ie. the probability of y @ receiver, when x is sent.

 

Bayes theorem

 

[math]P(x|y) = P(y|x)\frac{P(x)}{P(y)}[/math]

 

and

 

[math]P(x,y) = P(x|y)P(y)[/math]

 

Shannons reasoning was that, given a channel characterized by P(y|x), there is a marginal distribution P(x) that maximizes the information divergence between the joint distrubution P(x,y) and the indepdenent case P(x)P(y).

 

Shannon limit [math] C = max_{P(x)} S_{KL}(P(x,y)||P(x)P(y)) [/math]

 

So the shannon capacity is 0 iff x and y are independent.

 

[math]S_{KL}[/math] is the Kullback–Leibler divergence, http://en.wikipedia.org/wiki/Kullback-Leibler_divergence

 

Anyway the key is that the shannon assumes the channel is known, perfectly known. Which is not always a realistic scenario.

 

/Fredrik

Link to comment
Share on other sites

I have posted many thought provoking ideas on this site hoping that I might find some others that also find these things interesting. but nobody seems to be interested in having any intelligent conversation. everyone is only interested in selling their own personal pet idea and in tearing apart everyone elses ideas. this is whats wrong with, not just this site but, the scientific establishment and indeed the whole world as far as I can tell.

 

How else could anyone tell if an idea had merit, other than by testing to see if it fails?

Link to comment
Share on other sites

threatening to excommunicate me is hardly a threat. I could care less.
This much is obvious in your lack of rigor.

 

I have the impression that no matter how simple and obvious something I say is that its going to be endlessly picked apart by at least one of you. and if you cant argue with what I said then you just pretend I said something else and argue with that.
Please point such instances out, as that would be Strawmanning and we give infractions to people who persist in fallacious logic like that.

 

I can get a little cranky sometimes and I probably shouldnt post when I'm in such a mood. I apologize for that.
Well done, sir. I think all here have fallen prey to this particular human frailty at some point. Apology accepted.

 

I have posted many thought provoking ideas on this site hoping that I might find some others that also find these things interesting. but nobody seems to be interested in having any intelligent conversation.
We are here to have scientific conversations. Intelligent conversations, as you have evidenced here, are at the whim of your definition of intelligent, which seems biased against scientific rigor.

 

everyone is only interested in selling their own personal pet idea and in tearing apart everyone elses ideas.
Please identify the kettle you are speaking of, granpa pot.

 

this is whats wrong with, not just this site but, the scientific establishment and indeed the whole world as far as I can tell.
You see a flaw but instead of learning whether you're right you simply point a finger and scream. This is neither science nor good learning.

 

if you excommunicate me then so much the better.
I agree. Your bubble seems comfortable to you and it is cold and harsh outside of it.
Link to comment
Share on other sites

I have the impression that no matter how simple and obvious something I say is that its going to be endlessly picked apart by at least one of you.

Yup :-D

 

This is because we discuss (most) things here with rigour and a certain enthusiasm for detail. You just need to factor that in when you are deciding what to post. Some people find this is too much effort and leave, but is that our loss? Those who do stay usually rapidly improve their critical and communication skills.

 

and if you cant argue with what I said then you just pretend I said something else and argue with that.

Use of logical fallacies is against the rules, so feel free to report posts which you think qualify.

Link to comment
Share on other sites

I'll try to get this topic back on track -- to wit, how are scientific theories produced, and can the process be mechanized.

 

The problem of forming a scientific law (a simple empirical equation) is hard all by itself. Suppose we have 1000 measurements of some phenomenon. What are the independent variables? Suppose we randomly pick one. One can then easily construct an exact fit by fitting those 1000 measurements against a 999 degree polynomial. The result will of course be absolutely meaningless. Even empirical regression is a bit of an art. Why a polynomial rather than say an exponential? This variable as the independent variable rather some other variable, or some other variables? What constitutes error in the fit versus error in the model?

 

If generating a good regression model is a challenging task, coming up with a good theory that explains why the model is the way it is is extremely hard. Everyday (non-genius) scientists do a passable job at regression on a regular basis. They don't do such a good job at putting the pieces together; that takes true genius.

 

An example is that of black body radiation.

 

Wien's Law for black body radiation (1986):

 

[math]I(\nu,T) = a \nu^3 \frac 1 {e^{\frac {b\nu}{T}}}[/math]

 

where a and b are constants of proportionality. What made Wien pick that particular model? It was purely ad-hoc, not motivated by any extant theory. This model yields a good match to the high frequency spectrum of a black body radiator, but not the low frequency spectrum. (Not surprising: Wien was trying to model the high frequency spectrum.)

 

Planck saw that a simple correction would yield a model that matched the full spectrum of a black body radiator:

 

[math]I(\nu,T) = a \nu^3 \frac 1 {e^{\frac {b\nu}{T}}-1}[/math]

 

Seeing that this seemingly simple change would work for the full spectrum was a small stroke of genius. It was however a purely empirical relationship, with two tuning parameters. A little more work reduced this to one tuning parameter, h (Planck's constant):

 

[math]I(\nu,T)=\frac {2h\nu^3}{c^2}\frac 1 {e^{\frac {h\nu}{kT}}-1}[/math]

 

The other constants are k, Boltzmann's constant, and c, the speed of light. Seeing that Planck's constant has much wider applicability ([math]E=h\nu[/math]) than an empirical tuning parameter in the black body radiation law took several more years.

 

 

Edit:

I found an interesting read on the above topic: http://physicsworld.com/cws/article/print/373

Link to comment
Share on other sites

Hi again :)

 

I had said that I wouldn't enter the forum again , but I just couldn't resist. I think that I was treated with dogmatism, but I think I will give it another try.

 

First of all, I apologize in advance for my English, I know it is not very good, but at least I am trying :)

 

Second, I would like to underline that I do not know much about physics. I have read a lot of papers, books, etc but "Shannon's theorem" etc are all greek to me (that's odd, since I am Greek :D )

On the other hand, I know much more things about epistemology and my original thread referred to an epistemological issue.

 

My point was that science has not yet established even one of the ways (methods) we use in order to produce theories. As a consequence, we do not know what to change if things don't go well and we can't produce well-founded theories.

 

I will get straight to the point: Kuhn, Lakatos and many other philosophers of science/ scientists have observed that each normal science (according to Kuhn "normal science takes place in the context of a shared paradigm. It consists of three main scientific activities: "determination of significant fact, matching fact with theory, and articulation of theory.... Work under the paradigm can be conducted in no other way, and to desert the paradigm is to cease practicing the science it defines."- http://www.arps.org/users/hs/kochn/QuantitativeReasoning/Glossary.html)

 

is founded on paradigm (Lakatos called it "research program", I call it "basic truth").

 

They have also observed that this paradigm (or fundamental belief, dogma, doctrine, etc) determines everything that the scientist who has accepted it thinks of or does. In other words, the effectiveness of a scientist in his scientific work is directly proportional to the quality of his paradigm.

 

Despite its "omnipresence", philosophers say that it is impossible to locate it and express it verbally. On the other hand, they claim that an improvement of the prevailing paradigm is equal to a scientific revolution. Indeed, we can notice that if we take a look at the development of physics. The latter has made a huge stride when Einstein changed the prevailing paradigm that had been introduced by Newton. (of course, there are many other examples from the history of all sciences)

 

The question is: How does this paradigm determine the scientific (both theoretical and research) work? How can we locate it and express its content? Can we improve a factor that we do not know? Should we just wait for scientific revolutions to happen?? In other words, in which way do we produce theories and how can we improve this way, so that we interpret the phenomena we study?

 

Effie

 

ps. I think (hope) that you justify the fact that in my post I don't give an answer to all the previous posts, don't you? :embarass:

Link to comment
Share on other sites

Kuhn was right in that scientists do belong to specialized communities and that scientific revolutions do occur. He was, IMHO, flat-out wrong regarding incommensurability and incommensurable paradigms.

 

Kuhn's first transition is from pre-scientific beliefs to science. Those initial scientific concepts are not so much incommensurable with the pre-scientific beliefs as they are incompatible with them. Galileo and Darwin were not attacked because their ideas were incommensurable with the pre-scientific beliefs. They were attacked because their ideas were commensurable with the pre-scientific beliefs -- and in utter conflict with those beliefs. The attacks on Galileo and Darwin resulted because their pre-scientific critics knew exactly what they were saying.

 

Kuhn asserts in The Structure of Scientific Revolutions that incommensurability between the pre- and post-scientific revolution communities results in part because of shifts in the meanings of key terms. Kuhn specifically picks mass as an example:

… the physical referents of these Einsteinian concepts are by no means identical with those of the Newtonian concepts that bear the same name. (Newtonian mass is conserved; Einsteinian is convertible with energy. Only at low relative velocities may the two be measured in the same way, and even then they must not be conceived to be the same.)

This is an out-and-out straw man, and a rather deliberate one at that (Kuhn was educated as a physicist). The concept of rest mass in relativity is completely commensurable with mass in classical physics.

 

The scientific revolutions in physics at the onset of the 20th century did not create incommensurable communities. The tools, techniques, and lingo of the new physics were comprehensible by physicists whose training was purely Newtonian. Most older physicists joined these new communities.

 

Some physicists admittedly did choose not to join the relativistic and quantum mechanics communities. Einstein, for example, never accepted the uncertain nature of quantum mechanics. There was no incommensurability, however. Einstein understood quite well the tools, techniques, and lingo of quantum mechanics. He used this understanding to develop the EPR paradox. The EPR paradox has since been resolved and has been observed. The EPR paradox is a paradox in the Greek sense (a rather counterintuitive result) but is not a paradox in the logical sense (a contradiction that disproves something).

Link to comment
Share on other sites

It is easier to talk about how theory is created, if you create a theory and watch how it occurs. This requires self awareness. Before you finalize your ideas, use some self observation to see how the process actually works within your own head. This is called first hand data collection since theory forms in the human brain and not outside the human brain. You can't expect to know why the car is making smoke if you don't look under the hood. The brain is a physical tool and not just an abstraction.

 

The analogy is putting together a play. The audience sees opening night with everything 90% there, but doesn't see the uncoordinated theatre company that begins, going through all the growing pains as they rehearse and make a lot of mistakes. One can get the impression over night they have this well organized performance. All the mistakes before opening night of a theory come from hunches, reason and the unexpected. But these will not make it into open night if you can avoid it. The audience won't clap.

 

Let us try an experiment to collect first hand data for the creation of theory. Let us blue sky theorize what will be the next major genetic change in humans. There is no right answer, so you can say what you want. Don't just recite what someone else has placed in the final fabrication stage of opening night. Use your own imagination and you will see the process with first hand observation. Theorists need thick skin because all theories will get crapped on until you get past the inception stage. There is a tendency to nip it in the bud whether it is right or wrong. That negativity is part of theory creation since it affects the direction one is allowed to lead the theory down. Or what the audience expects to see on opening night.

 

Let me give an example. The earth has an iron core. This is not based on direct data but inference. One is not suppose to say this but I like reality. If you try to point any theory away from this unproven expectation, forget it. It would be treated like an off broad way play. Most people will stick to Broadway. The best theorists know the needs of the critics up front. During the second half you may need to weed just to please the critics. He likes high kicks in the dance numbers, so you may need to add some even if it is not want you really wish to do. There is a art to science theory.

Link to comment
Share on other sites

My point was that science has not yet established even one of the ways (methods) we use in order to produce theories. As a consequence, we do not know what to change if things don't go well and we can't produce well-founded theories.

 

From my point of interest here, the quest for knowledge certainly goes hand in hand with a quest for the method of acquiring knowledge. Neither knowledge, nor the method is perfect, nor do I expect a day when it is.

 

But through history, human knowledge as well as human method for knowledge has developed. Long time ago there was no developed scientific methods at all and not much of a rational system to establish what is truth. It evolved from faith, and appeal to authority(probable opinion historically referred to is opinion of authority), to more objective ideals of establishing truth. However there's more to the method than to secure objectivity, also efficiency is an issue.

 

Should we just wait for scientific revolutions to happen??

 

I think we work as fast as we can, given the contstrained resources, but if you ask if there is a way to speed time? :) I don't think so.

 

If you're ask if your or me have to wait for the mainstream ideas to be revolutionized before thinking outside box, I think the obvious answer is no.

 

In other words, in which way do we produce theories and how can we improve this way, so that we interpret the phenomena we study?

 

This is similar to question I ask myself too. Since we are dealing with somewhat creative processes here, I think this question is related to artificial intelligence, and self-evolving learning models. This is why I think evolutionary models is needed in physics, that applies also to the notion of law.

 

Relating to for example the discussion http://www.scienceforums.net/forum/showthread.php?p=455842#post455842

 

My associations to this entire discussion is not as much historical as that of the view of physics and physical law as a result of an evolution. To try to describe this evolution, to the extent possible (to perfectly do so, is I think impossible), is strongly related to the question of a self-organising and observer, existing in an a priori unknown environment.

 

My personal starting points is closely related to statistical inference, but from a point of view that is more intrinsic, that respects the constraints, in particular information capacity constraints of an observer. There is only a limited set of inferences possible from the inside. I think this explains certain properties, and possible makes predictions about self-organisation.

 

/Fredrik

Link to comment
Share on other sites

Kuhn asserts in The Structure of Scientific Revolutions that incommensurability between the pre- and post-scientific revolution communities results in part because of shifts in the meanings of key terms.

 

I think that the shift in the meanings of terms is the outcome of the paradigm shift and not the opposite. Consequently, incommensurability begins from the paradigm level and is extented to terms, methods, researches, etc. However, from what I have read on the subject, I agree that "the concept of rest mass in relativity is completely commensurable with mass in classical physics".

 

Einstein, for example, never accepted the uncertain nature of quantum mechanics.

 

Thus the famous quote: "Quantum mechanics is certainly imposing. But an inner voice tells me that it is not yet the real thing. The theory says a lot, but does not really bring us any closer to the secret of the 'old one'. I, at any rate, am convinced that He does not throw dice"

The above only proves that physics (as any other science) hasn't yet completed its assignment. In other words, we can expect that more revolutions will take place.

After all, we all know what has happened since Lord Kelvin claimed that physics has almost completed its work, except for some details that remain to be arranged :)

 

It is easier to talk about how theory is created, if you create a theory and watch how it occurs.

 

I totally agree. Besides, that's what I have been supporting since the beginning: we do know how theories are produced, this cannot be diputed, given that we do produce them. The question is: can we express this knowledge verbally?

 

 

 

All the mistakes before opening night of a theory come from hunches, reason and the unexpected. But these will not make it into open night if you can avoid it. The audience won't clap.

 

All I am asking is: where do this "hunches" come from? Could Newton have had "hunches" regarding gravity if he hadn't suspected its presence? That is to say, we can form opinions only regarding factors we know or we think that exist. Molecular biologists for example, due to their current paradigm, seem to have no " hunches" regarding the role of the endogenous fields of cells, despite the fact that their presence has been verified beyond any doubt.

 

 

 

If you try to point any theory away from this unproven expectation, forget it. It would be treated like an off broad way play. Most people will stick to Broadway. The best theorists know the needs of the critics up front.

 

If I have perceived correctly what you have written, I have to say that I agree. After all, science must first rely on the knowledge it already has. The "known" is the most solid, secure substrate on which we can rely in order to take a look at the unknown. Every normal science refuses to discuss ill-founded ideas which are not supported by well established, valid data and knowledge. That's only natural and essential for its normal function. Nevertheless, science should test all the alternative options. If we refuse to do this and we remain stuck to our old beliefs, exactly how scientific is our approach? Let me remind you that all the knowledge we today consider obvious were mere uncertain opinions when they were presented for the very first time.

 

"The philosophy of one century is the common sense of the next." Henry Ward Beecher

 

From my point of interest here, the quest for knowledge certainly goes hand in hand with a quest for the method of acquiring knowledge. Neither knowledge, nor the method is perfect, nor do I expect a day when it is.

 

Yeah, that's true. Universe (reality) has always proved itself far more complex than our most elaborated opinion/theory etc and we cannot expect that this will change, at least soon.

 

It evolved from faith, and appeal to authority(probable opinion historically referred to is opinion of authority), to more objective ideals of establishing truth.

 

Let me disagree a bit. Today there are sciences which remain religiously stuck to their basic principles, even if they have been superseded by research data. Molecular biology and psychiatry are the most apparent examples. Of course, research is objective and produces valid data, which unfortunately cannot be utilized.

 

 

If you're ask if your or me have to wait for the mainstream ideas to be revolutionized before thinking outside box, I think the obvious answer is no.

 

In a previous post I had written a quote from Aristotle to which nobody paid any attention, because "Aristotle has expressed many wrong ideas". (as far as I know, Planck has supported some wrong ideas, too. E.g. at first denied the wave/ particle duality introduced by Einstein, but nobody has dicarded his other ideas due to this "misfortune") Anyway, according to Aristotle, nobody should enter a normal science without having expressed basic questions regarding its basic principles/ axioms, because his thought "will be tied". According to Kuhn, scientists- due to the way in which they are educated- learn to serve the prevailing paradigm of their science.

 

Once in the box, it is very hard to succeed in thinking outside box, for many reasons which we could discuss about.

 

 

My personal starting points is closely related to statistical inference, but from a point of view that is more intrinsic, that respects the constraints, in particular information capacity constraints of an observer. There is only a limited set of inferences possible from the inside. I think this explains certain properties, and possible makes predictions about self-organisation.

 

I would like to learn more about this approach

 

Effie

 

ps thank you very much for the link :)

Link to comment
Share on other sites

The question is: can we express this knowledge verbally?

Yes, but not succinctly. This is part of the reason why people have to go to school for many, many years.

 

You asked a related question in your first post: can the process be mechanized? The answer is no. We can't even fully mechanize (describe algorithmically) the process of creating scientific laws, simple empirical relationships. This is exactly what the no-free-lunch theorems say.

 

Wikipedia article: No free lunch in search and optimization

 

The papers that started it all:

Wolpert, D.H. (1996) "The lack of a priori distinctions between learning algorithms," Neural Computation, 8(7), 1341–1390. (Note: PDF file starts at the end. Go to the last page and read backwards.)

 

Wolpert, D.H. and MacReady, W.G. (1997) "No free lunch theorems for optimization," Evolutionary Computation, 1(1), 67–82.

 

A bibliography page: http://www.no-free-lunch.org

 

If merely producing good scientific laws is not fully mechanizable, what hope can there be for mechanizing the process of explaining scientific laws (i.e., coming up with a theory)?

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.