Does Real Science Need To Be Falsifiable?

The Pixie

Well-known member
This is an issue that impacts on the evolution/creation debate, so while that is the background, I would like to focus on examples outside of that.

With that in mind, I will be basing this on an article, "Falsifiability and physics" in Symmetry Magazine a couple of years ago.

From it:

Popper wrote in his classic book The Logic of Scientific Discovery that a theory that cannot be proven false—that is, a theory flexible enough to encompass every possible experimental outcome—is scientifically useless. He wrote that a scientific idea must contain the key to its own downfall: It must make predictions that can be tested and, if those predictions are proven false, the theory must be jettisoned.

There are, I think, two things to consider here. The first is science as a corpus of knowledge, the second is science as a process. The former is what mankind broadly believes as the facts of science. Relativity, for example, is broadly accepted and part of mainstream science. It is taught in schools as a part of the body of knowledge. For a hypothesis to become a part of that in my view, it has to be falsifiable. Can anyone think of an example where that is not true?

What scientists do in labs is the process of the science. Should that be falsifiable? No. Much of it is speculative, or just do an experiment and see what happens. The process of arriving at a good scientific hypothesis really does not matter, as long as it works. If you end up with a hypothesis that is falsifiable (and its predictions tested), that is what matters.

But where does this falsifiability requirement leave certain areas of theoretical physics? String theory, for example, involves physics on extremely small length scales unreachable by any foreseeable experiment. Cosmic inflation, a theory that explains much about the properties of the observable universe, may itself be untestable through direct observations. Some critics believe these theories are unfalsifiable and, for that reason, are of dubious scientific value.

String theory, as far as I know, is speculative. It is not mainsteam science, it is a hypothesis (or several!) that physicists are considering but are not (as a group) committed to.

With regards to cosmic inflation, Wiki says that actually predictions have been confirmed - if theory was not confirmed surely the theory would have been falsified? I am not a physicist, and stand to be corrected here.

Tracy Slatyer of MIT agrees, and argues that stringently worrying about falsification can prevent new ideas from germinating, stifling creativity. “In theoretical physics, the vast majority of all the ideas you ever work on are going to be wrong,” she says. “They may be interesting ideas, they may be beautiful ideas, they may be gorgeous structures that are simply not realized in our universe.”

A lot of science is about speculation. Speculation is great, and concerns about falsifiability really should not be an issue at this stage. But if you want me to accept your hypothesis as mainstream science, that is when you need to show it is falsifiable. Or if you prefer, that is when you have to show that it makes predictions, and those predictions were shown to be good.

Beyond falsifiability of dark matter or SUSY, physicists are motivated by more mundane concerns. “Even if these individual scenarios are in principle falsifiable, how much money would [it] take and how much time would it take?” Slatyer says. In other words, rather than try to demonstrate or rule out SUSY as a whole, physicists focus on particle experiments that can be performed within a certain number of budgetary cycles. It's not romantic, but it's true nevertheless.

SUSY is supersymmetry, by the way.

Speculation about dark matter and SUSY is great, but they are not mainstream science, and they will not become mainstream science until the predictions they make have been tested and confirmed. There are practical concerns, and if you lack the budget to test the predictions of your theory then tough, it is not going to become part of mainstream science.

General relativity also makes predictions about things that are untestable by definition, like how particles move inside the event horizon of a black hole: No information about these trajectories can be determined by experiment.

General relativity makes a lot of good predictions, and that is enough to make it science. The idea that a theory has to be tested in every extreme is nonsense. Newton's laws of motion also make predictions about how matter behaves in a black hole, and is, undoubtedly, wrong.

Worth noting, in fact, also that it is still good science if a hypothesis' predictions are right most of the time. We still teach Newton's laws of motion at school because the predictions are mostly right. What makes Newton's laws science is that they make firm predictions about what we should observe, and those predictions turn out to be correct most of the time.

A hypothesis that makes no predictions, or that makes untestable predictions, or that makes predictions that are wrong as often as they are right, is not fit to become part of the corpus of knowledge of science.
 

TeabagSalad

Well-known member
I studied physics at university (many, many years ago), what you are saying is correct. If a hypothesis doesn't have a way of being tested and falsified then it shouldn't be part of the body of scientific knowledge.

I believe that cosmic inflation has largely been shown to occur. However, it does lead to a load more questions for which we currently don't have answers for - and may not be able to answer in the foreseeable future.

String Theory is something that I don't subscribe to, because of the reasons you have outlined - with our current level of technology it isn't something that we can test. That having been said, I have a friend who I met at university; he is a far better physicist that I could ever have become and went on to have a career in physics research. He "loves" String Theory, it's sort of like a hobby for him. He knows that it can't be tested and may never be able to be, but he sees it as useful. He once compared his study and work on the theory as being like a journalist or novelist doing a crossword - it doesn't directly help them with their work or serve a really useful function (apart from being a bit of fun). But, in the same way that a crossword may teach a journalist a new word or phrase, String Theory may teach a physicist a new mathematical technique which may benefit their actual work.

It's part of the fun of science - using your imagination to come up with ideas and techniques, which may not be useful now, but who knows what such ideas may reveal in the future.
 

Komodo

Well-known member
[. . .] What scientists do in labs is the process of the science. Should that be falsifiable? No. Much of it is speculative, or just do an experiment and see what happens. The process of arriving at a good scientific hypothesis really does not matter, as long as it works. If you end up with a hypothesis that is falsifiable (and its predictions tested), that is what matters.

But where does this falsifiability requirement leave certain areas of theoretical physics? String theory, for example, involves physics on extremely small length scales unreachable by any foreseeable experiment. Cosmic inflation, a theory that explains much about the properties of the observable universe, may itself be untestable through direct observations. Some critics believe these theories are unfalsifiable and, for that reason, are of dubious scientific value.

String theory, as far as I know, is speculative. It is not mainsteam science, it is a hypothesis (or several!) that physicists are considering but are not (as a group) committed to....
I'd be a little looser with my taxonomy than that, for however microscopically little that's worth. String theory, so far as I know, is not done in labs, but more or less on paper. But so was both Special Relativity and General Relativity, initially. Do we want to say that Einstein wasn't doing science, that it only became science with the experimental confirmation? I guess we could say that if we wanted, but it sounds strange to me to say that Einstein wasn't doing science, just some fancy mathematics which happened to turn out to have scientific value. So I would also not want to say that String Theory isn't science; at the very least, it's bounded by the body of scientific knowledge we do possess, which to my mind sets it apart from pure mathematics or metaphysics. That is, if it were shown that String Theory, or one variant thereof, necessarily implied that molecules couldn't exist, or that stars couldn't form, etc., then that theory would be falsified. I agree that it isn't scientific knowledge as we generally understand it, and may never be, but that might just be an unfortunate result of our lack of resources.

(By the way, would Einstein be the last Hall of Fame scientist who never did an experiment or even designed an experiment, but worked entirely with pencil and paper?)
 
Last edited:

The Pixie

Well-known member
I'd be a little looser with my taxonomy than that, for however microscopically little that's worth. String theory, so far as I know, is not done in labs, but more or less on paper. But so was both Special Relativity and General Relativity, initially. Do we want to say that Einstein wasn't doing science, that it only became science with the experimental confirmation? I guess we could say that if we wanted, but it sounds strange to me to say that Einstein wasn't doing science, just some fancy mathematics which happened to turn out to have scientific value. So I would also not want to say that String Theory isn't science; at the very least, it's bounded by the body of scientific knowledge we do possess, which to my mind sets it apart from pure mathematics or metaphysics. That is, if it were shown that String Theory, or one variant thereof, necessarily implied that molecules couldn't exist, or that stars couldn't form, etc., then that theory would be falsified. I agree that it isn't scientific knowledge as we generally understand it, and may never be, but that might just be an unfortunate result of our lack of resources.

(By the way, would Einstein be the last Hall of Fame scientist who never did an experiment or even designed an experiment, but worked entirely with pencil and paper?)
But that is my point. He was doing science the process, but what he proposed was not part of the corpus of scientific knowledge until it was falsifiable.
 

Komodo

Well-known member
But that is my point. He was doing science the process, but what he proposed was not part of the corpus of scientific knowledge until it was falsifiable.
Ah, OK. Sorry about that; I thought by saying String Theory wasn't mainstream science, you were sort of banishing it to the realm of math or metaphysics. I agree of course that it isn't established scientific knowledge.
 

inertia

Super Member
This is an issue that impacts on the evolution/creation debate, so while that is the background, I would like to focus on examples outside of that.

With that in mind, I will be basing this on an article, "Falsifiability and physics" in Symmetry Magazine a couple of years ago.

From it:

Popper wrote in his classic book The Logic of Scientific Discovery that a theory that cannot be proven false—that is, a theory flexible enough to encompass every possible experimental outcome—is scientifically useless. He wrote that a scientific idea must contain the key to its own downfall: It must make predictions that can be tested and, if those predictions are proven false, the theory must be jettisoned.

There are, I think, two things to consider here. The first is science as a corpus of knowledge, the second is science as a process. The former is what mankind broadly believes as the facts of science. Relativity, for example, is broadly accepted and part of mainstream science. It is taught in schools as a part of the body of knowledge. For a hypothesis to become a part of that in my view, it has to be falsifiable. Can anyone think of an example where that is not true?

I can't think of any example where this is not true. In general, scientists worldwide understand that falsification is an integral part of scientific methodology. A hypothesis, no matter how elegant it may be, is ultimately judged with cold-hard experimental measurements. This is where tangible access and mathematical prediction provide insight into the world around us.

Relativity is considered the second pillar of physics for very good reasons.

What scientists do in labs is the process of the science. Should that be falsifiable? No.

I disagree.

Experimental physics requires measurements and measurements require an understanding of their uncertainty, precision, and accuracy. As in theory, it relies on mathematics to provide a coherent framework. If a measurement is not repeatable, that in itself should be a warning to the experimentalist that something is very wrong and an example of how falsification is used in the laboratories.

While replying to this part of your OP, I am reminded of two well-educated scientists who published a paper on an experiment showing cold fusion at room temperature in 1989. As it turned out, they failed to take into account the background radiation in their laboratory. Their experiment was falsified. ( a real-time embarrassment too )

Much of it is speculative, or just do an experiment and see what happens. The process of arriving at a good scientific hypothesis really does not matter, as long as it works. If you end up with a hypothesis that is falsifiable (and its predictions tested), that is what matters.

But where does this falsifiability requirement leave certain areas of theoretical physics? String theory, for example, involves physics on extremely small length scales unreachable by any foreseeable experiment. Cosmic inflation, a theory that explains much about the properties of the observable universe, may itself be untestable through direct observations. Some critics believe these theories are unfalsifiable and, for that reason, are of dubious scientific value.

String theory, as far as I know, is speculative. It is not mainsteam science, it is a hypothesis (or several!) that physicists are considering but are not (as a group) committed to.

String theory is mainstream in the sense that it provides many, many predictions. However, since it cannot be falsified, as far as I am aware, it remains in the realm of scientific philosophy for the time being. According to Brian Greene, "a theory's success can be used as an after-the-fact justification for its architecture, even when that architecture remains beyond our ability to access directly*".

With regards to cosmic inflation, Wiki says that actually predictions have been confirmed - if theory was not confirmed surely the theory would have been falsified? I am not a physicist, and stand to be corrected here.

Tracy Slatyer of MIT agrees, and argues that stringently worrying about falsification can prevent new ideas from germinating, stifling creativity. “In theoretical physics, the vast majority of all the ideas you ever work on are going to be wrong,” she says. “They may be interesting ideas, they may be beautiful ideas, they may be gorgeous structures that are simply not realized in our universe.”

A lot of science is about speculation. Speculation is great, and concerns about falsifiability really should not be an issue at this stage. But if you want me to accept your hypothesis as mainstream science, that is when you need to show it is falsifiable. Or if you prefer, that is when you have to show that it makes predictions, and those predictions were shown to be good.

What interests me currently here is our ability to falsify the origins of cosmological time. Since the acceleration four-dimensional spacetime includes time itself, time too should have inflated. The hope is that our ability to detect and measure primordial gravitational waves using LIGO will provide the resolution necessary to measure newly created time, locally, with colliding black holes or neutron stars. An increasing delay should be detectable in theory between the predicted and observed signals.

Wait-and-see...

Beyond falsifiability of dark matter or SUSY, physicists are motivated by more mundane concerns. “Even if these individual scenarios are in principle falsifiable, how much money would [it] take and how much time would it take?” Slatyer says. In other words, rather than try to demonstrate or rule out SUSY as a whole, physicists focus on particle experiments that can be performed within a certain number of budgetary cycles. It's not romantic, but it's true nevertheless.

SUSY is supersymmetry, by the way.
Speculation about dark matter and SUSY is great, but they are not mainstream science, and they will not become mainstream science until the predictions they make have been tested and confirmed. There are practical concerns, and if you lack the budget to test the predictions of your theory then tough, it is not going to become part of mainstream science.
...

Well, it's quite possible that all of the mathematical symmetries within the supersymmetric standard model are not actually realized. Still, particle interactions only remain consistent if their parameters are ultra finely tuned. By "finely tuned" we mean a single part in a million billion. With supersymmetry, bosons (particles with whole spin numbers), and fermions (particles with half-integer spins) provide canceling quantum mechanical contributions, and the theory ensures that substantial cancelations are actually attained.

Historically, Glashow, Salam, and Weinberg experimentally established that in very high energy environments the electromagnetic and weak forces actually unite based on a Grand Unification hypothesis. They received the Nobel Prize for their contribution to physics for their discovery and demonstrated a more symmetric union of forces in an early universe state. Since that historic experiment, other experimentalists have refined the tests with the strong, weak, and electromagnetic forces demonstrating a near convergence with tiny discrepancies. When supersymmetry is incorporated, the discrepancies vanish.

Physicists continue to explore GUT and SUSY hypotheses.



* Brian Greene, "The Hidden Reality parallel universes and the deep laws of the cosmos", First Vintage Books edition, Nov 2011, page 193
 
Last edited:

The Pixie

Well-known member
I can't think of any example where this is not true. In general, scientists worldwide understand that falsification is an integral part of scientific methodology. A hypothesis, no matter how elegant it may be, is ultimately judged with cold-hard experimental measurements. This is where tangible access and mathematical prediction provide insight into the world around us.

Relativity is considered the second pillar of physics for very good reasons..
Okay

I disagree.

Experimental physics requires measurements and measurements require an understanding of their uncertainty, precision, and accuracy. As in theory, it relies on mathematics to provide a coherent framework. If a measurement is not repeatable, that in itself should be a warning to the experimentalist that something is very wrong and an example of how falsification is used in the laboratories.
What about an experiment to see what happens with no hypothesis in mind? Is that not science then?

String theory is mainstream in the sense that it provides many, many predictions. However, since it cannot be falsified, as far as I am aware, it remains in the realm of scientific philosophy for the time being. According to Brian Greene, "a theory's success can be used as an after-the-fact justification for its architecture, even when that architecture remains beyond our ability to access directly*".
So do you think the development of string theory is NOT part of the process of science? Given your previous paragraph, that would seem to be the case, but this seems less clear.

What interests me currently here is our ability to falsify the origins of cosmological time. Since the acceleration four-dimensional spacetime includes time itself, time too should have inflated. The hope is that our ability to detect and measure primordial gravitational waves using LIGO will provide the resolution necessary to measure newly created time, locally, with colliding black holes or neutron stars. An increasing delay should be detectable in theory between the predicted and observed signals.

Wait-and-see...
That would seem a great example of something that is not currently falsifiable, so not part of the body of knowledge, but it is - in my view - part of the process.

Well, it's quite possible that all of the mathematical symmetries within the supersymmetric standard model are not actually realized. Still, particle interactions only remain consistent if their parameters are ultra finely tuned. By "finely tuned" we mean a single part in a million billion. With supersymmetry, bosons (particles with whole spin numbers), and fermions (particles with half-integer spins) provide canceling quantum mechanical contributions, and the theory ensures that substantial cancelations are actually attained.
Again, to me that is part of the process of science.
 

inertia

Super Member
What about an experiment to see what happens with no hypothesis in mind? Is that not science then?

I'll use an example to illustrate the concept.

Engineers perform validation tests and utilize well-understood scientific concepts to demonstrate the performance of instruments. Since they are not exploring any specific physical phenomenon, and no specific hypothesis is being evaluated, this process is classified as engineering.

That said, serendipity and thought experiments occur within and outside laboratories. Hypotheses are generated and exploration begins. In physics, mathematics is the primary tool used to develop models.

So do you think the development of string theory is NOT part of the process of science? Given your previous paragraph, that would seem to be the case, but this seems less clear.

Not at all. It is a scientific model that requires instruments with ultra-high precision to validate. Until validation occurs, it remains in the category of scientific philosophy. Even so, models constitute the core of scientific exploration but we understand that the process is incomplete.

That would seem a great example of something that is not currently falsifiable, so not part of the body of knowledge, but it is - in my view - part of the process.

Again, to me that is part of the process of science.

Indeed.
 
Last edited:

The Pixie

Well-known member
I'll use an example to illustrate the concept.

Engineers perform validation tests and utilize well-understood scientific concepts to demonstrate the performance of instruments. Since they are not exploring any specific physical phenomenon, and no specific hypothesis is being evaluated, this process is classified as engineering.

That said, serendipity and thought experiments occur within and outside laboratories. Hypotheses are generated and exploration begins. In physics, mathematics is the primary tool used to develop models.
What is your point? I am not sure if you are saying this is science or not.

Not at all. It is a scientific model that requires instruments with ultra-high precision to validate. Until validation occurs, it remains in the category of scientific philosophy. Even so, models constitute the core of scientific exploration but we understand that the process is incomplete.
As far as I can tell, you are agreeing with me.
 

inertia

Super Member
What is your point? I am not sure if you are saying this is science or not.


As far as I can tell, you are agreeing with me.

There is a difference between applying scientific knowledge as a tool and doing science to explore. I used the field of engineering to demonstrate experiments (tests) that are performed without the need to validate a scientific hypothesis.

It was my reply to the following questions:
What about an experiment to see what happens with no hypothesis in mind? Is that not science then?

Both engineering and science use the process of making observations, gathering evidence, and documenting, but the disciplines are not identical. Engineering is not equal to science even though they frequently overlap skillsets. Sometimes their goals are identical precisely because of the need to provide instruments that might validate theoretical concepts.

Gathering evidence Documenting observations with control group.JPG


 
Last edited:

The Pixie

Well-known member
There is a difference between applying scientific knowledge as a tool and doing science to explore. I used the field of engineering to demonstrate experiments (tests) that are performed without the need to validate a scientific hypothesis.
Ah, okay. I fully agree with that. See my thread on SETI.
 
Top