Go Back   AFA Forums > Science, Logic and Reason > Ask a scientist

Ask a scientist Not up to speed with science jargon but have a sciency question?

Reply
 
Thread Tools
  #11  
Old 21st July 2016, 06:37 PM
MikeJay MikeJay is online now
Senior Member
 
Join Date: Jul 2011
Posts: 1,259
Default Re: Data vs theory

I've made the typical error of conflating postulate and theory! ...but I don't think I'm alone in this thread.
Reply With Quote
  #12  
Old 21st July 2016, 07:21 PM
Spearthrower Spearthrower is offline
AFA Member
 
Join Date: Oct 2015
Posts: 4,119
Default Re: Data vs theory

Quote:
MikeJay said View Post
So Tycho collected data for no purpose and Kepler then trawled through that data looking for patterns without first having any theories as an objective to trawl through the data in the first place?

Not really 'no purpose' - he did it because it excited him, because he wanted to shore up the divide between a Ptolemaic and Copernican understanding, and because it brought him renown.

But we have to remember, he thought the Sun orbited the Earth! This is not a man famed for his explanatory powers, but for his observation and record-keeping.



Quote:
MikeJay said View Post
I'm trying to come to grips with the notion of collecting or obtaining data for the sole purpose of then trying to find patterns to base a new theory on without having any preconceived idea of a theory to begin with.
For the sake of inquiry? For interest? I also find it difficult to understand your difficulty - imagine people collecting butterflies: they do it because they collect, because of the adventure, because butterflies are beautiful - but a theorist even generations later might be able to use their bean-counting to provide an actual insight into butterfly distribution, or migration, or extinction.

Finally, some people may collect data to support their idea, but then find the data doesn't fit.

Last edited by Spearthrower; 21st July 2016 at 07:24 PM.
Reply With Quote
Like odd liked this post
  #13  
Old 22nd July 2016, 11:01 AM
Stub King's Avatar
Stub King Stub King is offline
Take my advice, don't listen to me
 
Join Date: Jun 2015
Location: Melbourne
Posts: 802
Default Re: Data vs theory

I think that in this day and age it is a little bit of both. during the early years of experimental science there was no shortage of data but not many (as we now know in hindsight) good theories. moreover, the type of data of concern was widely and easily available and pretty evident: differences in species, planetary motion, chemical reactions. So it was more a case of finding a good theory that fits the data and then validating it. nowadays, I think it is a little more chicken vs. egg. case in point in string theory (which should really be called string hypothesis). sure you could say that the data is that we have 4 forces, a bunch of particles and there is obviously something that makes them work, so what could it be? but this isn't particularly useful because we have good theories for the "macro" data but as we delve into more and more specificity, the data may not be all that easy to find. nor can you collide particles in the LHC and ask "well, what theory can I build from that?". today, it is a co-dependent relationship where you have theoretical physicists (in this case) coming up with models and experimental physicists (in this case) trying to construct experiments that will validate (or not) the theory. so the experiment is guided by what the theory is and the results are used to refine the theory. sometimes, an experiment or an observation comes up with inexplicable or unexpected data requiring a brand new explanation.
__________________
The less people know, the more stubbornly they know it. (Osho)
Reply With Quote
  #14  
Old 24th July 2016, 08:46 PM
DanDare's Avatar
DanDare DanDare is offline
Religion or Reality, choose...
 
Join Date: Mar 2009
Location: Brisbane, Queensland, Australia
Posts: 7,393
Default Re: Data vs theory

It goes 'round and 'round. Data suggests a hypothesis. Hypothesis suggests experiments. Experiments gather new data to put up against the hypothesis. Hypothesis is reinforced or alterations made or new hypothesis required. Spin. Rinse. Repeat.

Sometimes a perceptual framework, like a hypothesis, can get you to see old data in a new light. Sometimes new data from casual observation, or experiment aimed at something else, boinks into an existing theory or hypothesis and prompts change.

The important thing is that thesis cannot override data.
__________________
"History, I believe, furnishes no example of a priest-ridden people maintaining a free civil government".
-Thomas Jefferson

Burden of proof is the obligation on somebody presenting a claim to provide evidence to support its truth (a warrant). Once evidence has been presented, it is up to any opposing "side" to show the evidence presented is not adequate. If claims were accepted without warrants, then every claim could simultaneously be claimed to be true.

History isn't written by the victors. It's written by the people with the time machines.
Reply With Quote
  #15  
Old 24th July 2016, 09:58 PM
hackenslash's Avatar
hackenslash hackenslash is offline
Trust me, I'm not a doctor.
 
Join Date: Mar 2012
Location: People's Republic of Mancunia, Antipodes
Posts: 1,537
Default Re: Data vs theory

Quote:
tmorg said View Post
Scientists should follow the data right and not the theory? This question was prompted by the book I am reading "The Elegant Universe: Superstrings, Hidden Dimensions and the Quest for the Ultimate Theory" [1999] by Brian Greene. Data first and then theory in that order? This is preferable to formulating a theory and then looking for the data to fit it e.g. superstrings? (I realise there is already a discussion about this.)
First of all, great book. One of my favourite works of popular science.

Here's the thing:

Theories are data led, but investigations are theory led. It's a complex relationship, with one leading to the other and back again all the time. In order to formulate a theory, we need data. In order to derive predictions about what observations and experiments will show, we need theory. In order to test theory, we need data. Rinse and repeat.

M-Theory is an interesting case, not because it's any different to other theories with regard to how it progresses, but because it's been around a long time without strongly testable predictions. That's not to say it hasn't made any, but that what predictions have arisen from it are not easily testable.

The big problem for M-Theory is its incompleteness. Ultimately, as the book will tell you, theorists don't actually know what the final theory will look like, because the mathematics driving it is so horrendously difficult, which has resulted in progress being slow.

There are those who think it should be abandoned and that too much time and money has been spent on it already. I'm not one of them. I do think that it's favoured above other lines of enquiry where it shouldn't be, but I also still think that it's our most promising route to resolving the central problem of physics. Other approaches look promising in terms of unifying gravity and quantum mechanics (and I read a new paper this week by Sean Carroll et al detailing another new approach that I think is incredibly elegant), but none of them have the complete scope of M-Theory. It's also the most elegant potential solution I've come across because, at bottom, there's a single entity, the brane (poly-dimensional string). This resolves an awful lot of what are seen as issues in the standard model, because we need to understand why6 there are so many different particles, especially when only a few of them are required to construct a universe like the one we observe. M-Theory is the only framework we have that offers an explanation for this 'particle zoo'.

We should be looking in other areas, but we should still investigate M-Theory. It's also worth noting that M-Theory has provided huge advances in mathematics generally, so it's fruitful in that respect, regardless of whether it's breaking enough ground in unification.

I'll link a couple of blog posts for you.

http://reciprocity-giving-something-...hy-art-of.html
http://reciprocity-giving-something-...ction-and.html
__________________

Reply With Quote
Like Spearthrower, Strato, DanDare liked this post
  #16  
Old 25th July 2016, 10:55 PM
Strato's Avatar
Strato Strato is offline
What Me Deluded?
 
Join Date: Jul 2012
Location: The Bellarine, Geelong.
Posts: 5,211
Default Re: Data vs theory

Loved your posts on the art of philosophy and how science proceeds, and on fallacies, linked to, hack.
__________________
DNA is a coded description of ancestral environments, a 'genetic book of the dead'. Science in the Soul: Selected Writings of a Passionate Rationalist, Richard Dawkins.
Reply With Quote
Like hackenslash liked this post
  #17  
Old 25th July 2016, 11:13 PM
hackenslash's Avatar
hackenslash hackenslash is offline
Trust me, I'm not a doctor.
 
Join Date: Mar 2012
Location: People's Republic of Mancunia, Antipodes
Posts: 1,537
Default Re: Data vs theory

Thanks. Number9?
__________________

Reply With Quote
  #18  
Old 26th July 2016, 11:54 AM
Darwinsbulldog's Avatar
Darwinsbulldog Darwinsbulldog is offline
AFA Member
 
Join Date: Jan 2009
Location: Perth
Posts: 18,198
Default Re: Data vs theory

Data, theory.....it is the what came first, chicken or egg conundrum. In the beginning there were gods, which were placeholders for shit we couldn't understand. Gods were the explaination [often not very good ones, but the only explanations on offer].
The empiriacal evidence drove descriptive and predictive models. For example, the explanation for why there are so many species of birds and how they change over time was due to Darwin's Theory.

But as Mayr & Diamond noted in their studies of New Guinea birds, the indigenous tribes already had an empirical theory of bird species that described and predicted behaviour, but did not explain. There was an almost exact correlation between indigenous and scientific bird species. [The "why" bit was explained by legends and gods].

Vaccuum energy is in a similar state today. No explaination, just an empirical model.

Although finding out the 'whys" in science is the goal of many, science is really about "the hows" -descriptive and predictive models. Even proximate causes can be hard to pin down, never mind ultimates causes. With the rise of super-gods [Yahweh & co], the assumption of ultimate causes became absurd. [The four "Omnis"]

God is of course, the ultimate null hypothesis, because such a universe would look nothing like our own, or might indeed be impossible.
__________________
Just stick to the idea that science tests falsifiable hypotheses to destruction.
Reply With Quote
  #19  
Old 26th July 2016, 07:16 PM
Stub King's Avatar
Stub King Stub King is offline
Take my advice, don't listen to me
 
Join Date: Jun 2015
Location: Melbourne
Posts: 802
Default Re: Data vs theory

Quote:
hackenslash said View Post
I read a new paper this week by Sean Carroll et al detailing another new approach that I think is incredibly elegant
any chance for a link?
__________________
The less people know, the more stubbornly they know it. (Osho)
Reply With Quote
  #20  
Old 26th July 2016, 08:05 PM
hackenslash's Avatar
hackenslash hackenslash is offline
Trust me, I'm not a doctor.
 
Join Date: Mar 2012
Location: People's Republic of Mancunia, Antipodes
Posts: 1,537
Default Re: Data vs theory

There's a link in this blog post explaining it.
__________________

Reply With Quote
Reply

Bookmarks

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +11. The time now is 05:07 AM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2017, vBulletin Solutions Inc.
Feedback Buttons provided by Advanced Post Thanks / Like (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.