importance of hypothesis

[xxx200]

hi

 

the question is why should we use a hypothesis? is it necessary to use hypothesis in every situation?

 

in ancient times natural philosophers use question-answer method to know nature. they do not use hypothesis. in early modern ages, baconian method ( sir francis bacon) were used by sir thomas brown, issac newton, john stuart mill etc. also emphasizes that no hypothesis should be used. sir issac newton said in his principia that "hypotheses non fingo" (I don't make hypotheses). in his optiks he said "hypotheses have no place in experimental science."

 

but in modern science hypothesis is the starting point of research. why it is so? who started it? why noted scientists such as issac newton were so against it?

 

please answer.

 

 

source: baconian method

Edited September 17, 2012 by xxx200

[Nobrainer]

In my opinion, original thought comes before hypothesis, it's more that a hypothesis, it is a swirling cloud with a sudden clear understanding .... And then maybe ...

[Bignose]

I frankly don't care what the starting point is: hypothesis, questions and answers, 'logic', tea leaves, or a blow to the head. The bigger point is that science to a large extent doesn't care about that first step. It cares about making predictions and then comparing that prediction to measurements.

 

Whatever that 1st step is, what is really important is to get past just that 1st step, and actually make predictions and check their accuracy.

 

Too, too many people think that 1st step is sufficient, that their creativity, logic, or tea leaves is enough to create something meaningful. And while there is a lot of good to be said for being creative and coming up with new ideas -- the crucible of science is almost completely: what predictions can be made using the idea, and how good are those predictions. We have moved past the dark ages where ideas are judged based upon how much one likes the idea. In science, ideas are judged by accuracy of predictions.

Edited September 17, 2012 by Bignose

[xxx200]

I frankly don't care what the starting point is: hypothesis, questions and answers, 'logic', tea leaves, or a blow to the head. The bigger point is that science to a large extent doesn't care about that first step. It cares about making predictions and then comparing that prediction to measurements.

 

Whatever that 1st step is, what is really important is to get past just that 1st step, and actually make predictions and check their accuracy.

 

Too, too many people think that 1st step is sufficient, that their creativity, logic, or tea leaves is enough to create something meaningful. And while there is a lot of good to be said for being creative and coming up with new ideas -- the crucible of science is almost completely: what predictions can be made using the idea, and how good are those predictions. We have moved past the dark ages where ideas are judged based upon how much one likes the idea. In science, ideas are judged by accuracy of predictions.

 

why we need prediction in studying nature? suppose we study how trees grow. should we first predict how trees grow and then measure the accuracy of the prediction? why not just simply observe how trees grow?

[uncool]

why we need prediction in studying nature? suppose we study how trees grow. should we first predict how trees grow and then measure the accuracy of the prediction? why not just simply observe how trees grow?

Because it has the tendency to lead to conclusions based on incomplete information.

 

For one of the huge examples, look up N-rays. Scientists at the time "simply observed" their existence without forming the necessary predictions. Had they made predictions on whether N-rays would be observed in different situations, they would have noticed no difference between whether the supposed causes of N-rays were there or not - and so they would have found out that N-rays are a trick of the mind.

=Uncool-

[xxx200]

Because it has the tendency to lead to conclusions based on incomplete information.

 

For one of the huge examples, look up N-rays. Scientists at the time "simply observed" their existence without forming the necessary predictions. Had they made predictions on whether N-rays would be observed in different situations, they would have noticed no difference between whether the supposed causes of N-rays were there or not - and so they would have found out that N-rays are a trick of the mind.

=Uncool-

 

so you said that to know something invisible or subtle prediction is needed. but why then sir issac newton said that prediction has no place in experimental science in his optiks.

[uncool]

so you said that to know something invisible or subtle prediction is needed.

No, that isn't what I said.

 

What I said is that lack of prediction has the tendency to lead to conclusions based on incomplete information.

but why then sir issac newton said that prediction has no place in experimental science in his optiks.

For one thing, because he lived 350 years ago - long before most scientific philosophy had been discussed. Most of the relevant philosophy got its start in the 19th century if I remember correctly.

 

The basic problem with a lack of predictions is that it can tend to lead to the Texas Sharpshooter fallacy: shoot at a wall, and then draw the target around the hole, and claim that you are a sharpshooter because you managed to hit the center of the target. The same thing can happen with theories - do the experiment, see the outcome, and then tailor your theory to that outcome, and claim that your theory is correct because it explains that outcome. If, instead, you predict rather than postdict, then it's like setting the target in place before shooting - if you still manage to hit the center of the target, then yes, you are most likely a sharpshooter.

=Uncool-

Edited September 17, 2012 by uncool

[Phi for All]

but why then sir issac newton said that prediction has no place in experimental science in his optiks.

He didn't. His comments were about hypotheses, not prediction, and it was about experimental philosophy, NOT experimental science. The full quote is, “For whatever is not deduced from the phenomena must be called a hypothesis; and hypotheses, whether metaphysical or physical, or based on occult qualities, or mechanical, have no place in experimental philosophy”.

[Bignose]

why we need prediction in studying nature? suppose we study how trees grow. should we first predict how trees grow and then measure the accuracy of the prediction? why not just simply observe how trees grow?

 

Quite simply, observation alone will only tell you about the things you have observed. The power of prediction is that you attempt to create knowledge about things you haven't directly observed.

 

For example, by making a prediction about how tall a tree is based on how many years its been alive, and then confirming that prediction using a suitable sample of experiments, you don't have to cut down every tree in the forest and count its rings to make a good guess about the distribution of ages of trees in that forest.

 

Or, to take a recent example, making predictions allowed us to land a probe on Mars without having had to go to Mars and experiment how to land probes on Mars.

[xxx200]

For one thing, because he lived 350 years ago - long before most scientific philosophy had been discussed. Most of the relevant philosophy got its start in the 19th century if I remember correctly.

 

but he discovered the foundation of modern science: gravity, light, motion etc. he must have some erudition in him.

 

The same thing can happen with theories - do the experiment, see the outcome, and then tailor your theory to that outcome, and claim that your theory is correct because it explains that outcome.

 

but how we can explain a thing without seeing it first? how could we even predict about a thing without seeing it properly? what we should predict?

 

If, instead, you predict rather than postdict, then it's like setting the target in place before shooting - if you still manage to hit the center of the target, then yes, you are most likely a sharpshooter.

 

no it is like setting the carriage before the horse rather than setting the horse before carriage. it is very stupid.

 

what is going on now is that people set a prediction and then do the experiment and interpret the outcome in a way to support the prediction. this practice gives birth to some of the most bizzare concepts which is beyond our wildest imagination.

 

i think newton is right.

=Uncool-

[uncool]

but he discovered the foundation of modern science: gravity, light, motion etc. he must have some erudition in him.

Just because he managed to do those things doesn't imply that he knew everything.

 

but how we can explain a thing without seeing it first? how could we even predict about a thing without seeing it properly? what we should predict?

You predict by noticing patterns. For example, Kepler predicted that planet's orbits will always be ellipses by seeing that the planets that he'd seen orbited in ellipses. And it turns out to be true - in general, planets orbit in ellipses. Newton predicted that gravity took the form F = G m_1 m_2/r^2 in general from Kepler's observations (plus some more), and Cavendish showed that those predictions were accurate on a small scale. Yes, observations come first, but the prediction step is crucial - and must come before experimentation.

 

no it is like setting the carriage before the horse rather than setting the horse before carriage. it is very stupid.

 

what is going on now is that people set a prediction and then do the experiment and interpret the outcome in a way to support the prediction.

Please support this assertion. The point of making a prediction is that the prediction sets very tight bounds on which outcomes are considered to support the prediction; what you're describing is specifically what doesn't happen.

this practice gives birth to some of the most bizzare concepts which is beyond our wildest imagination.

 

i think newton is right.

=Uncool-

You have based what you are saying on an assertion which is false.

=Uncool-

[Bignose]

Do you want a case in point where your observation --> theory pattern fails?

 

Take a look at your thread on matter. In post #12 (http://www.scienceforums.net/topic/68992-what-is-matter-anyway/page__view__findpost__p__702544)

 

you wrote:

 

so molecules consist of atoms. many atoms gather to create a molecule. ok? the nature of molecule must come from nature of atoms or the that of the bonds. solid atoms gather to create solid molecules. then these Solids, liquids, and gases are states of atoms too.

 

so by this logic there are 3 types of atom :Solids, liquids, and gases. now furthur info on atom:

 

And as I pointed out in that thread, this is just wrong. The atoms themselves don't change with the phases. Just the energy level by which the atoms and molecules interact -- that's why we have the gas, liquid, and solid phases, as well as several other phases of matter you didn't even bother to name.

 

You observed gases, liquids, and solids and you tried to extrapolate that back to the atoms (dare I say, predicted that those would also be properties of the atoms?). It is a good try, but it fails in the face of experimental evidence.

 

That's why the prediction & comparison step is the real meat of science today. It is no longer good enough to make statements that sound "good enough" or "logical enough" or "plausible enough".

 

this practice gives birth to some of the most bizzare concepts which is beyond our wildest imagination.

 

Lastly, I think it is fine to have concepts that are wild and imaginative. Science needs that. Craves it even. New thoughts and concepts are always desired. But, they are never in isolation. Just having a new thought or idea is not enough to be worth anything to science. To be worth something, you have to use that new idea to make predictions and then compare how good those predictions are to experiment. That's the worth.

 

New ideas are needed, because otherwise there would never be new predictions. But then those new predictions are judged based on how well they agree with reality. And then once we have valid predictions, we can do all sorts of things previously un-thought of. Like, land a rover on Mars using a parachute and reverse thrusters.

 

If we didn't have exceptionally well validated models (predictions) of celestial mechanics, gravity, booster rockets, atmospheric drag, and chemical reactions (just to name a few), we wouldn't have been able to make that landing on Mars successfully. Would you have been comfortable paying millions of dollars to land a probe on Mars with just "well, we watched a lot of things land here on Earth"?

[Phi for All]

i think newton is right.

=Uncool-

Why did you sign this post with someone else's username? That's fairly rude.

[uncool]

Why did you sign this post with someone else's username? That's fairly rude.

I think he simply forgot, when quoting my post, to put my signature in quotes.

=Uncool-

[Phi for All]

I think he simply forgot, when quoting my post, to put my signature in quotes.

=Uncool-

That makes sense. I guess assuming he was making fun of you is a bad hypothesis.

[xxx200]

You predict by noticing patterns. For example, Kepler predicted that planet's orbits will always be ellipses by seeing that the planets that he'd seen orbited in ellipses. And it turns out to be true - in general, planets orbit in ellipses. Newton predicted that gravity took the form F = G m_1 m_2/r^2 in general from Kepler's observations (plus some more), and Cavendish showed that those predictions were accurate on a small scale. Yes, observations come first, but the prediction step is crucial - and must come before experimentation.

 

 

well then the pattern of scientific investigation is : observation A, prediction B, observation B, prediction C......................................... observation N. ok. is modern scientific investigation goes on in this pattern?

[Ronald Hyde]

All this 'splainin' is too complicated for me. I make you a simpler explanation. A hypothesis is a concept introduced for the purpose of testing a theory.

When one tries to construct and test a theory, one 'entertains' a hypothesis or so. If the theory works and requires that the hypothesis be included for the

theory to work, then it no longer remains a hypothesis, it gets promoted to being a premiss of the theory. Big promotion. If it fails it is discarded, and

a new hypothesis is introduced and the theory retested. A hypothesis that covers a wide class of objects is called a conjecture.

 

Simpler, yes?

[Bignose]

well then the pattern of scientific investigation is : observation A, prediction B, observation B, prediction C......................................... observation N. ok. is modern scientific investigation goes on in this pattern?

 

More or less. Again, to me the starting point is rather arbitrary. Also, I do think it is important not to gloss over some things in the chain "prediction B, observation B, prediction C".

 

I would have written this is: prediction B, observation B, calculate error between prediction B and observation B, evaluate the feasibility or likelihood of the model that lead to prediction B based on that calculated error, reject the model or refine it as needed (maybe no refinement at all), THEN made prediction C and repeat.

[xxx200]

More or less. Again, to me the starting point is rather arbitrary. Also, I do think it is important not to gloss over some things in the chain "prediction B, observation B, prediction C".

 

I would have written this is: prediction B, observation B, calculate error between prediction B and observation B, evaluate the feasibility or likelihood of the model that lead to prediction B based on that calculated error, reject the model or refine it as needed (maybe no refinement at all), THEN made prediction C and repeat.

 

in science THERE IS NO PLACE FOR ARBITRARINESS. science is concrete, logical and based on sound facts. i really cannot appreciate arbitration in science.

Edited September 19, 2012 by xxx200

[Bignose]

in science THERE IS NO PLACE FOR ARBITRARINESS. science is concrete, logical and based on sound facts. i really cannot appreciate arbitration in science.

 

Well, there is a new word you haven't used in this thread before. What exactly do you mean by 'arbitration'? What exactly is wrong with what I posted? What are you objecting to here?

 

Edited to add -- I do apologize. I thought this was a reaction to the methodology. It is in fact a reaction to my assertion that 'the starting place' is rather arbitrary. I wanted to leave the above note in there since I had it posted for several hours.

 

And then I am going to go ahead and defer to the next post in this thread by Phi For All, and his defense of the arbitrary starting point.

 

Again, sorry I misinterpreted your post here.

Edited September 19, 2012 by Bignose

[Phi for All]

in science THERE IS NO PLACE FOR ARBITRARINESS. science is concrete, logical and based on sound facts. i really cannot appreciate arbitration in science.

There are multiple ways to conduct valid experiments that produce data that can lead you to conclusions about your hypothesis. The methodology must be sound and that's why it's discussed among peers and tested by those peers using similar sound methodology. Where you start is arbitrary, but that doesn't mean the methodology used is. And "facts" are not as important to scientific methodology as evidence that supports or refutes.

 

Btw, arbitration ≠ an act of arbitrariness.

[dmaiski]

a hypothesis is, at its most basic, a theory

it is setting the framework around which you do your experiment

 

an example:

i believe pig's can fly

to confirm this hypothesis i will observe pigs, throw them off high buildings and see if they can fly

 

a null hypothesis, is a morons idea

it helps outline conditions for your experiments failure, but at the most fundamental level it is a redundant step

especially if your experiment partially confirms your hypothesis

 

i also believe it was shown to be a fallacy a while back

a null hypothesis is only useful for statistical modelling, not actual experimentation, and thus should only be stated in the results section when you are doing some form of quantitative statistical analysis of the obtained results, and should be specific to those statistical tests on the variable that you are measuring in the test

 

 

 

Edited September 19, 2012 by dmaiski

[xxx200]

Well, there is a new word you haven't used in this thread before. What exactly do you mean by 'arbitration'? What exactly is wrong with what I posted? What are you objecting to here?

 

Edited to add -- I do apologize. I thought this was a reaction to the methodology. It is in fact a reaction to my assertion that 'the starting place' is rather arbitrary. I wanted to leave the above note in there since I had it posted for several hours.

 

And then I am going to go ahead and defer to the next post in this thread by Phi For All, and his defense of the arbitrary starting point.

 

Again, sorry I misinterpreted your post here.

 

you said that prediction is arbitrary . it means that prediction is not based on sound logic and fact. this is the meaning of arbitraryness. prediction must be based on sound logic. arbitrary prediction is like a sand castle. it can be broken down easily.

 

if you have to predict, predict based on careful observation.

[Phi for All]

you said that prediction is arbitrary . it means that prediction is not based on sound logic and fact. this is the meaning of arbitraryness. prediction must be based on sound logic. arbitrary prediction is like a sand castle. it can be broken down easily.

 

if you have to predict, predict based on careful observation.

Are you purposely misunderstanding what's being written? That's the second time you've rewritten what someone else wrote.

 

Bignose wrote, "Again, to me the starting point [of a scientific investigation] is rather arbitrary". HE DIDN'T SAY PREDICTION IS ARBITRARY.

 

 

 

Perhaps your signature should read, "ask the question, get the answer, and then tailor your response to that answer by strawmanning it, and then claim you're correct because you refuted the wrong answer."

[Ronald Hyde]

I'll give an example that might help understanding a bit. Let's suppose you entertain the hypothesis that results are independent of the method

used to obtain them, i.e. the way that you measure something should not affect the results. Sounds perfectly logical, right? But if you try to

apply it in real life, at the quantum level, it falls flat on it's face. Some people will say that the quantum rules only apply at the small scale, but this

is an old fallacy that should be put to rest. You can describe the Rainbow with QM, and the very first rule of QM, is that you formulate the problem

from the beginning of the experiment. from the beginning to the end. So in the case of the rainbow, the photon comes all the way from the Sun and

ends up at your eye. What if you substitute other instruments for the eye, or modify the eye by putting sunglasses on. You have entertained a new

hypothesis and redefined the experiment. So by the nature of the experiment you entertain a hypothesis, whether you think you are or not. so every

experiment, no matter how large or small, entertains a number of hypothesis. Some people will say 'that problem only needs elementary QM to explain

the experiment, or they will say it needs advanced QM. That is a fallacy too. There is only one QM, which just has this little handful of rules. The

elementary or advanced part is entirely in the proper formulation, and always uses the same little set of rules. But people often skip the

'formulate the complete experiment' part and that is where errors creep in.