Falsifiability - has it been fully discredited?

  • Thread starter Nereid
  • Start date
In summary: The third problem is that confirmationism is vulnerable to 'just-so stories': tales of how some confirmation could have come about that are not actually logically related to the theory under consideration. So for example, say you have a theory that the world is round. You might be able to find some observations that seem to support this theory, but you might also find observations that seem to disprove it. Now, if you were a confirmationist, you would say that the theory is confirmed because both sets of observations are related to the postulates of the theory - in other words, because they are consistent with the theory. But this is not actually the case - the observations that seem to
  • #1
Nereid
Staff Emeritus
Science Advisor
Gold Member
3,401
3
"Falsifiability" - has it been fully discredited?

Karl Popper introduced the concept of 'falsifiability' as the touchstone of (modern) science. In essence, Popper claimed that scientists propose theories and (other) scientists seek (experimental, observational) evidence to 'falsify' these theories. Once 'falsified' by a (good) experimental or observational result, the (now falsified) theory is abandoned by (all) scientists, and new theories proposed. And so on ... (of course I'm simplifying, but I hope I've not done too much damage to Popper's idea).

Critics of Popper, including Kuhn and Lakatos, fairly quickly showed that this is not at all how science actually works. For example, good observations of the position of the planet Uranus showed it was not at all where it 'should' be, according to Newtonian (gravitational) theory. Rather than 'abandon' Newton, scientists worked on a 'patch' - don't abandon Newton, look for something that will make Newton whole! Years (how many? I should look it up) later, Neptune was discovered, and the anomaly in Uranus' orbit went away.

Many years later, a similar situation arose regarding the orbit of Mercury. Again, well after the 'anomaly' was independently verified, and the Newtonian predictions clearly seen to be way beyond the 'error bars' of the observations, scientists didn't regard Newton as 'falsified' ... until a new theory came along!

In the history books, it's all clear-cut; but if one doesn't have the benefit of hindsight, then what scientists actually do when faced with good observational/experimental results which, prima facie 'falsifies' a theory is quite different from Popper's description.

Modern examples: neutrino flavours (now more or less 'resolved'), dark matter in spiral galaxies (lively debate), the Pioneer anomaly (a true anomaly - the data are good enough to 'falsify' GR/Newton, there are no 'patches', yet no one has 'abandoned' GR!).

Curiously, while those who study science have long since put 'falsifiability' into the file 'good idea at the time, didn't pan out', a great many scientists appear to have embraced the idea! At least, they use the term freely when talking about what they do, to non-scientists. Any ideas why this is?

But, I've written this from memory, and haven't even checked Popper, let alone his critics. How bad is my memory? If my memory is pretty good, how inaccurate is my summary?
 
Physics news on Phys.org
  • #2
Nereid said:
Karl Popper introduced the concept of 'falsifiability' as the touchstone of (modern) science. In essence, Popper claimed that scientists propose theories and (other) scientists seek (experimental, observational) evidence to 'falsify' these theories. Once 'falsified' by a (good) experimental or observational result, the (now falsified) theory is abandoned by (all) scientists, and new theories proposed. And so on ... (of course I'm simplifying, but I hope I've not done too much damage to Popper's idea).

The idea is that falsified theories must either be modified or abandoned. The basic logic behind it is that scientific theories cannot be proven true, but they can be proven false. So he considers those theories "scientific" that make risky claims. Then if falsifying evidence comes in, the theory is known to be false, by modus tollens.

I'm going to jump down to the bottom first, then address the stuff in the middle.

Curiously, while those who study science have long since put 'falsifiability' into the file 'good idea at the time, didn't pan out', a great many scientists appear to have embraced the idea! At least, they use the term freely when talking about what they do, to non-scientists. Any ideas why this is?

Yes, it's because the kernel of falsificationism goes hand-in-hand with inductive reasoning, which science (being an a posteriori discipline) relies on. To really appreciate falsificationism you have to look at what it replaced: confirmationism, which says that theories should be accepted as scientific if they are confirmed by experimental evidence. Confirmationism has the following three difficulties.

First, it allows tautological theories to be considered scientific. But as every student of logic knows a tautology is not contingent. So since it is known to be true a priori, without any need for experimentation, what point is there in saying that it is scientific?

To illustrate the second (more serious) problem, start with a theory T, which has as its postulates a set of n statements {Ti}n. Confirmationism holds that T is confirmed not only if any of the Ti are confirmed, but also if any statement that can be deduced from the Ti are confirmed. But this goes against the very deductive logic that is relied on to obtain the set {Di} of deductions from {Ti}n. The syllogism described above can be represented symbolically as:

[itex]\bigwedge_{i=1}^nT_i\longrightarrow D_j[/itex]

[itex]D_j[/itex]

[itex]\therefore\bigwedge_{i=1}^nT_i[/itex],

is deductively invalid. In fact this fallacy is so well known that it has a name: "fallacy of affirming the consequent".

Finally, the third problem with confirmationism is that it holds that evidence that confirms any statement of a theory (be it a postulate or deduction) also confirms any logically equivalent statement. This problem is higlighted in Hempel's Paradox of the Ravens. The statement "All ravens are black" is clearly contingent. However the statement "All nonblack things are not ravens" is logically equivalent to our original statement. That means that finding, say, a white napkin must necessarily be construed as evidence that supports the statement that all ravens are black! But if you are a falsifcationist, you would not waste your time looking for white napkins to support your theory. You would be looking for a white raven.

Critics of Popper, including Kuhn and Lakatos, fairly quickly showed that this is not at all how science actually works.

Eh? Wasn't SR developed because of the falsification of Galilean spacetime with electrodynamic experiments? Wasn't QM developed because of the falsification of the classical model of radiation by observed blackbody spectra? Weren't the seeds of QED sown by the falsification of classical electrodynamics by the photoelectric effect?

Furthermore, what positive claims did these people offer up? What do they think is "the logic of scientific discovery"? See, you can say that Popper has been discredited, but even if I were to agree with you it still begs the question, "Whaddya got instead?"

For example,

(snip)

In the history books, it's all clear-cut; but if one doesn't have the benefit of hindsight, then what scientists actually do when faced with good observational/experimental results which, prima facie 'falsifies' a theory is quite different from Popper's description.

First of all, I'll note here that the examples that I snipped are astronomical examples. Astronomers do not, and can not, perform controlled experiments on the heavenly bodies. As I remarked at the beginning, falsificationism works because modus tollens (shown below) is a valid syllogism:

[itex]p \longrightarrow q[/itex]

[itex]\neg q[/itex]

[itex]\therefore \neg p[/itex].

In the examples you cite above, it is simply the case that if one would have concluded [itex]\neg q[/itex], one would have been mistaken. The fact that it may take a great deal of time and effort with many controlled experiments to establish [itex]\neg q[/itex] (and thus falsify p) does not in any way discredit the adoption of this logic to science, which I think is eminently reasonable.

There are ways to attack falsificationism. But I think that citing historical examples is not one of them. Falsificationism is purported to solve the problems associated with confirmationism, and to solve the problem of demarcation. The way to attack falsificationism is to show that it fails on one or both of those counts.

But no matter how well Popper's version is trounced, I fail to see how the kernel of his ideas could ever be dispensed with in science. What sense would it make to not reject an hypothesis in the face of incontrovertible evidence that is contrary to it?

Modern examples: neutrino flavours (now more or less 'resolved'), dark matter in spiral galaxies (lively debate), the Pioneer anomaly (a true anomaly - the data are good enough to 'falsify' GR/Newton, there are no 'patches', yet no one has 'abandoned' GR!).

Again, this has to do with whether a theory is really falsified, as opposed to apparently so. But I assure you that once a theory has been definitively shown wrong, it is certainly abandoned by those who do not find it useful (that is, by those who work in the domain in which the old theory is known to be wrong). As a graduate student I did research with the medium energy particle physics group, and if anyone were to ask, "Why don't you use nonrelativistic, classical scattering theory?" the immediate reply would have been, "Because nonrelativistic, classical scattering theory is wrong."
 
Last edited:
  • #3
One other point:

Nereid, you said in the Mentor's Private Forum that Popper's ideas have been "falsified". That is quite impossible, since Popper's falsificationism is prescriptive. That is, Popper's falsificationism defines what it means for a theory to be "scientific". If scientists do not choose to formulate theories that are falsifiable, and if they do not choose to engage in science by "conjecture and refutation", then it does not imply that Popper's ideas have been proven false, it simply means that scientists are not being "scientific" according to a strictly Popperian view. But proving Popper's definition of "scientific" false is impossible. Proving it unpopular or not useful, well that's another matter entirely...

Tom Mattson said:
There are ways to attack falsificationism. But I think that citing historical examples is not one of them. Falsificationism is purported to solve the problems associated with confirmationism, and to solve the problem of demarcation. The way to attack falsificationism is to show that it fails on one or both of those counts.

I really think this deserves a second read. Falsificationism is a philosophy, not a science. It cannot be overturned by empirical evidence or by the historical record. If the theory is to really be refuted, then the refutation must include how Popper failed to solve the problems he claimed to solve.

Our man Hugo Holbling chose precisely that route in his criticism of Popper.

http://www.galilean-library.org/falsificationism.html [Broken]
 
Last edited by a moderator:
  • #4
Excellent thread, too bad to use the
in the Mentor's Private Forum
for those interesting things.

I have a few more pennies for what it's worth. I think the scientific method in any of it's forms is almost sacred, which would elevate it to dogma, which is wrong though, there should not be dogma’s. Consequently, any scientific method would subject to falsifiability otherwise it would not be scientific.

But what does it matter in the quest for the truth. It's all about the end result, not about how to get there. You can stumble upon it by accident or work towards with thousands of scientists and never even get close.

Let's get a little more practical.

How about the street is wet, hence it rains (affirming the consequence). No, streets can get wet in a number of ways.

How about global warming. Greenhouse gasses cause warming. There is warming, and there are more greenhouse gasses, hence the warming is caused by greenhouse gasses. Irrefutable logic? No, because the quantitative factor is missing. The warming part attributable to greenhouse gasses may be 100%, 10%, 1%, 0.1% or whatever.

The Younger Dryas was a brief cold interruption of the warming after the last glacial maximum of the ice age. Right? Forget it, no way, the result of very sloppy thinking caused by a number of logical errors together forming a slippery slope. Yet there is not a single quaternary scientific soul who is even beginning to consider the extreme remote possibility that 'the real story' is completely different, despite all the "remarkable", "non understood" phenomena that engulfs this era.
Just an example to illustrate that. So what is needed to falsify the Younger Dryas being cold?

Then, the pole shift. Ever since the notion grew that in northernmost Siberia horses, aurochses, antelopes and lions roamed the grassy steppes, way above the arctic circle, whilst simultaneously, massive ice sheets covered Canada and the Northern USA, some 30,000 years ago (abundant evidence available), the obvious hypothesis was a “pole shift” and Canada being on the North pole 30,000 years ago.

What would a logical thinking scientist do with that idea? There is a tough nut to crack. How to even begin to apply a scientific method on that phenomenon?

Anyway, that question has emerged a number of times in the last century. Why is it so obvious that the standard solution has been repeatedly: a pole shift is impossible, so that did not happen. period, don’t ever mention it again. But that does not make the evidence go away, on the contrary, it continues to pile and not only in North Siberia. However, when another crackpot layman sticks up his finger and asks about a pole shift once more, he is kicked back in his corner with increasing strong force. But is the “pole shift” really and definitely falsified? What is impossible? Do we really understand all the processes in that gigantic ball under our feet, when we have to admit to encounter one surprise after the other?

What is needed to pose a credible “pole shift” hypothesis and have it follow the scientific method?
 
Last edited:
  • #5
Tom Mattson said:
. . . Popper's falsificationism defines what it means for a theory to be "scientific". If scientists do not choose to formulate theories that are falsifiable, and if they do not choose to engage in science by "conjecture and refutation", then it does not imply that Popper's ideas have been proven false, it simply means that scientists are not being "scientific" according to a strictly Popperian view. But proving Popper's definition of "scientific" false is impossible. Proving it unpopular or not useful, well that's another matter entirely...

Insightful comments. I've wanted to understand something like what Nereid's point seems to be, which is if during actual research anybody is thinking about falsification, or are they mainly either looking to find if some predicted condition/thing exists, or testing to see what (unpredicted) might happen.

My own in-progress concept has been, when scientists get together to discuss a theory, that is when falsification most likely arises because it's a theorizing technique. But during experimentation, it seems (from accounts I've read of how research has proceeded that've led to discoveries), the focus is more like detective work, searching for anything that helps explain how something known works or which reveals how something new might be made to work. How's that concept?
 
  • #6
Les Sleeth said:
But during experimentation, it seems (from accounts I've read of how research has proceeded that've led to discoveries), the focus is more like detective work, searching for anything that helps explain how something known works or which reveals how something new might be made to work. How's that concept?
As an experimentalist studying a phenomenon that has no explanantion yet, and aware of several other experiments of this nature, I'd say that the above statement is a mischaracterization of typical experimental physics...though it may apply somewhat to certain areas of high energy/particle physics. But then what do you mean by "something new"? Often, this 'something new' gets discovered by experiment, and further experiments are needed to better characterize/understand the phenomenon (this before any reasonable theory exists).

In short, experimentation serves not merely the purpose of verification/falsification, but is extremely productive in terms of unearthing new science. But that does not mean that experiments do not (and have not ) provide(d) several instances of falsification.
 
Last edited:
  • #7
Falsification does apply to those astronomical examples you gave, and it does "falsify" the hypothesis that they are obeying Newton's theory of gravitation AND that we have all the relevant data governign the motion. thus it doesn't falsify Newton; it states either Newton is wrong or we have incorrect data. or we are you saying that that the corrected model for uranus uses a different model for gravity?
 
  • #8
Les Sleeth said:
I've wanted to understand something like what Nereid's point seems to be, which is if during actual research anybody is thinking about falsification, or are they mainly either looking to find if some predicted condition/thing exists, or testing to see what (unpredicted) might happen.

It depends on the discipline, I think. In 4 years of college and 4 years of grad school, I had not heard of the term "falsification" even once. I had not heard it until Another God (who is a molecular bio major) mentioned it. I think it might be the case that scientists whose discipline is more qualitative than quantitiative necessarily rely on logical methodologies such as Popper's. At least they do so more than, say physicists, who have the logical buttress of mathematics that gives their theories the muscle of deductive validity. Furthermore all of the (mathematical) reasoning that physicists do has as a end product a number, which is precisely what is the end product of a measurement.

Something I've been wondering about myself is whether physicists even need to have the term "falsifiable" in their vocabularies to do their job properly, because falsifiability seems to be automatically built into quantitative theories.

My own in-progress concept has been, when scientists get together to discuss a theory, that is when falsification most likely arises because it's a theorizing technique.

The term "falsifiable" explicitly applies to theories, so yes I would agree.

But during experimentation, it seems (from accounts I've read of how research has proceeded that've led to discoveries), the focus is more like detective work, searching for anything that helps explain how something known works or which reveals how something new might be made to work. How's that concept?

Hmmm..."Explanations" are the stuff of theories, not experiments. I'd say that there are two main roles to experimentation. In areas in which theory leads experiment, the experimenters are testing the predictions of a theory. They are told what to expect under a precisely controlled circumstances, and they set up the experiment exactly as prescribed to see if what the theorists say is in fact so. And in areas in which experiment leads theory, the role of experimentation is to probe new ground and hopefully expose enough of it for theorists to form a cogent account of.
 
  • #9
Tom Mattson said:
They are told what to expect under a precisely controlled circumstances, and they set up the experiment exactly as prescribed to see if what the theorists say is in fact so.
Not to nitpick, but the underlined portion is not a true reflection of the way experiments are done. Theorists will rarely prescribe to an experimentalist, how s/he should set up an experiment - mostly because they can't. Experimental physics is not just manual labor or a process of following instructions, even when attempting to verify/falsify.
 
  • #10
Of course, the theorist will not tell the experimentalist where to put his detectors, how to build them, how to analyze his data, etc. What a theorist in, say, subatomic physics (the area with which I am most familiar) will say is something like, "We expect that you will detect the Primakoff effect in eta-prime production off nucleons in the forward direction, when the center of mass energy is much larger than all the other energy scales in the system," (a true statement in medium energy physics).

Naturally the devil is in the details, and this is where the experimentalist must use his skills.
 
  • #11
Naturally, I'm sticking up for the experimentalists here, but I'm also trying to raise what I believe is a pertinent (although hardly central) point to the discussion. The link between theory and experiment is often not straightforward. The seemingly simple question "but are you really measuring the effect that I'm talking about ?" (asked by a theorist of an experimentalist) is often highly non-trivial. In other words, interpreting what exactly it is that is being measured/detected in an experiment can be a matter of prolonged debate and even "controversy".

For example, look into the recent claim by Moses Chan, that his group has observed superfluidity in solid He-4 and the subsequent refutation by Ceperley and others, that they (Chan & Kim) see what they claim they see.

http://guava.physics.uiuc.edu/~nigel/courses/463/essays_f2004/files/roy.pdf [Broken]
 
Last edited by a moderator:
  • #12
Nereid said:
Karl Popper introduced the concept of 'falsifiability' as the touchstone of (modern) science. In essence, Popper claimed that scientists propose theories and (other) scientists seek (experimental, observational) evidence to 'falsify' these theories. Once 'falsified' by a (good) experimental or observational result, the (now falsified) theory is abandoned by (all) scientists, and new theories proposed. And so on ... (of course I'm simplifying, but I hope I've not done too much damage to Popper's idea).

Critics of Popper, including Kuhn and Lakatos, fairly quickly showed that this is not at all how science actually works. For example, good observations of the position of the planet Uranus showed it was not at all where it 'should' be, according to Newtonian (gravitational) theory. Rather than 'abandon' Newton, scientists worked on a 'patch' - don't abandon Newton, look for something that will make Newton whole! Years (how many? I should look it up) later, Neptune was discovered, and the anomaly in Uranus' orbit went away.

Many years later, a similar situation arose regarding the orbit of Mercury. Again, well after the 'anomaly' was independently verified, and the Newtonian predictions clearly seen to be way beyond the 'error bars' of the observations, scientists didn't regard Newton as 'falsified' ... until a new theory came along!

In the history books, it's all clear-cut; but if one doesn't have the benefit of hindsight, then what scientists actually do when faced with good observational/experimental results which, prima facie 'falsifies' a theory is quite different from Popper's description.

Modern examples: neutrino flavours (now more or less 'resolved'), dark matter in spiral galaxies (lively debate), the Pioneer anomaly (a true anomaly - the data are good enough to 'falsify' GR/Newton, there are no 'patches', yet no one has 'abandoned' GR!).

Curiously, while those who study science have long since put 'falsifiability' into the file 'good idea at the time, didn't pan out', a great many scientists appear to have embraced the idea! At least, they use the term freely when talking about what they do, to non-scientists. Any ideas why this is?

But, I've written this from memory, and haven't even checked Popper, let alone his critics. How bad is my memory? If my memory is pretty good, how inaccurate is my summary?


Popper never rejects Duhem's thesis that only sets of assumptions (theories) are falsifiable, he actually stresses that this is the case; nonetheless there is no fatal problem here (what matters is that the new conjectures must always provide extra novel predictions apart from acommodating 'puzzles' and direct anomalies).

The fact that scientists usually 'cling' to the old paradigm is not a real problem in his view, for example the Uranus/Neptune 'problem' can still be fully accommodated within popperianism (a minor auxiliary hypothesis is changed leading to a new falsifiable theory, making also additional novel predictions, but leaving Newtonian mechanics untouched).

Popper does not recommend what conjectures should be developped during periods of crises, what counts is to be 'bold' (provide extra novel predictions) and 'coming from simple and powerful principles'. For example, looking retroactively, geocentrism and the assumption that the celestial bodies move only in circles would have been discarded for the theory becomes more and more cumbersome (far from making things simpler) needing basically daily adjustments and without providing the required extra novel predictions.

Also it is entirely acceptable to attack the validity of the methodologies used to conduct the experiments (if there is reasonable justification) or to try to 'reduce' some existing 'puzzles' within the existing paradigm (no new premises added/dropped).

Moreover there are many situations when all serious scientists agree that some experiments are crucial, corroborating/falsifying even singular statements (e.g Rutherford's experiment confirming the existence of atoms, the rejection of locality in Aspect's experiment and so on). Sure some mention further underdetermination and theory ladenness but they are are far from discrediting the popperian falsificationism.

For example in the problem of Kaufmann's experiment Planck showed clearly that Einstein's theory should be adopted using concepts accepted by both the SR interpretation of Kaufmann's experiment (coming along with Lorentz model of the electron) and Kaufmann's own interpretation of the experiment (coming along with Abraham's model of the electron), confirming thus SR.

There are some problems of course (for example why theories should be simpler, why believe there is only one correct theory of everything, scientists seems rather interested to prove that a theory is more probable to be true etc) but popperian falsificationism is still workable and easy to use (and, as far as I know, so far no account of scientific practice is perfect). What is really put under doubt is the logical positivist account of science, basically discarded.

Probably this is why popperian falsificationism is still so popular among scientists, in spite of Feynman's observation that 'philosophy of science is about as useful for scientists as ornithology is for birds' (not at all the best of views, I suppose scientists are still rational beings, something which cannot be said of birds... :-) ).

This is not to say that mere popperianism is the best account of scientific practice, no. One of the most important problems with falsificationism I see is that it, like verificationism before, seems to exclude some scientific laws/theoretical constructs from science. If, as Popper himself says, we can refute only systems of statements then what is the status of single physical laws, absolutely necessary for the observed empirical success of a certain (wider) theory, but which are not falsifiable in isolation?

For example if 'a' is such a statement, not falsifiable in isolation (but absolutely necessary to explain the observed empirical success of a theory) and 'b','c','d'...'p' the rest of the set of statements (let's say all falsifiable in isolation or 'confirmed' by previous experiments) in the premises of that theory from which we can deduce a set of novel predictions 1,2,3...(some of them already 'confirmed' experimentally) it results, strictly following Popper, that we can say only that the set 'a' & 'b' &...& 'p' is falsifiable therefore scientific and that 'b', 'c', 'd', ...'p' are also scientific, part of science. But what is the status of 'a' taken in isolation (during the time when the theory it is part of is not disproved)? Is it part of science or not?

A good example can be found if we examine Newton's laws. Here the law of inertia is such a law (basically unfalsifiable in isolation) but which proves to be very fecund toghether with the other postualtes of Newtonian mechanics. What is its status, when taken in isolation? No clear answer can be drawn from Popper's view, though clearly the 'le bon sens' indicate that the law of inertia is a scientific law. One solution is to say that they are metaphysical, themata etc but clearly this is not acceptable within mere popperianism which intended to make a clear demarcation between metaphysics and science.

Yet despite these problems I maintain that an evolution of Popper's falsificationism is the best approach in the light, especially, of the problem of empirical basis (the inductive approach does favorize dogmatism), Feyerabend let the door too open to relativism etc. Personally I lean towards a post-lakatosian methodology, something ressembling Maxwell's 'metaphysical' falsificationism.
 
Last edited:
  • #13
Tom Mattson said:
Eh? Wasn't SR developed because of the falsification of Galilean spacetime with electrodynamic experiments? Wasn't QM developed because of the falsification of the classical model of radiation by observed blackbody spectra? Weren't the seeds of QED sown by the falsification of classical electrodynamics by the photoelectric effect?

Kuhn is not saying falsification does not or should not happen; he is attacking a
simplistic
model of falsification in which one single countervailing datum
brings a theory down. Sometime a patching a theory is the right way
to go, as in the Uranus example, and sometimes it gets ridiculous,
as in the famous example of epicycles.

Furthermore, what positive claims did these people offer up? What do they think is "the logic of scientific discovery"? See, you can say that Popper has been discredited, but even if I were to agree with you it still begs the question, "Whaddya got instead?"

Kuhn claims that science works principly by 'paradigms' and 'exemplars'
rather than explicit rules. Problems are solved by discovering similarities to
previously solved problems.
 
Last edited by a moderator:
  • #14
Few comments to my last post regarding Popper's view. Einstein said once that a paradigm shift occurs when enough 'problems' ('puzzles') accumulate in the old paradigm or when we invent a theory which not only accommodates all known facts but which is also capable to unify facts thought previously as being unconnected. Well it has to be said that Popper's view handle pretty well both cases.

There are the following possibilities:

1. The existing paradigm has some, few at the beginning, 'puzzles' (e.g. new 'facts', which seem to belong to the 'domain of definition' of the existing paradigm are discovered, but which do not seem to directly contradict the paradigm). And it has to be said that basically there are very few paradigms in practice without 'puzzles' and problems which could qualify as anomalies.

2. The interpretation of one or more experiments seems to falsify the existing paradigm, at least one of its novel predictions (direct 'anomalies' exist).

3. A new theory, unifying facts thought previously as being unconnected, is invented though the existing paradigm has very few (or none) 'anomalies'.


Attitudes of scientists in the case 1:

1.1 Scientists try to 'reduce' the puzzle/puzzles to the existing paradigm, no new auxiliary premise is added/discarded (as during 'normal science' periods).

1.2 Scientists devise hypotheses where some new auxiliary premises are added so that all previous facts are accommodated, the puzzles are accommodated, clear novel predictions are made.

1.3 Scientists devise theories where some old premises are dropped, some new postulates/auxiliary premises are added so that all previous facts + as many as possible of the existing 'puzzles' are accommodated, novel predictions are made.


In the case 2:

2.1 The interpretation of the experiment(s) seemingly falsifying the existing paradigm is attacked with pertinent justification (as did Planck in the case of Kaufmann's interpretation).

2.2 The existence of an anomaly is accepted. Some new auxiliary premises are added so that all previous facts are accommodated, the existing anomalies and as many as possible of the 'puzzles' are accommodated, clear novel predictions are made.

2.3 The existence of an anomaly is accepted. Some old premises are dropped, some new postulates/auxiliary premises are added so that all previous facts + all anomalies and as many as possible of the existing puzzles are accommodated, novel predictions are made.


In the case 3:

The unified theory is adopted as the new paradigm even if it contradicts an important part of the assumptions of the old paradigm.


Popper does not exclude any approach in the periods of 'crises', all what is required for the new paradigm is to be 'bold' (to make novel predictions apart from accomodating the anomalies, as many as possible), to 'come from unitary, powerful, principles' and to be stable on long run. Thus his view gives a pretty good idea of scientific practice though, as I've argued before, it is not the best account of scientific practice we have now.
 
Last edited:
  • #15
"Anomoly only appears against the background provided by the paradigm. The more precise and far-reaching the paradigm is, the more sensitive an indicator it provides of anomoly, and hence of an occasion of paradigm change"


SSR, p65
 
  • #16
Great discussion - far better than I expected! Thanks everyone, and especially Tom.

If I may clear a little underbrush away before trying to see the woods for the trees?

'discredited' vs 'falsified': in that other place I was being a little cheeky Tom; here I chose a slightly less naughty word.

Philosophy vs something else: the thing which interests me is the extent to which 'science' can be used to look at 'science' - and it's philosophers like Popper, Kuhn (is he really a philosopher?), and Lakatos (yes, he's my hero) who seem to me to have done the best job.

Why? Because 'it works!' You are reading some (English) text on a screen (moniter, whatever); you have never seen me (nor me you), you have likely no inputs about anyone else who has posted to this thread other than similar words. Yet we have all spent time and effort on this. How did this come about? (if the internet etc isn't sufficiently powerful, try your favourite example from modern medicine, or the equipment in your lab, or ...) However it came about, I want more of it, and I want to know how it 'works' and how we can make it 'work' better.

OK, so now we can see a mostly vertical, tall, solid-looking thing or three.

If 'science' is 'what scientist do' (let's use this as a starting point), then 'what DO scientist do?' Popper gave it a shot, and his 'falsification' idea got developed rather nicely (thanks Hugo, 'sophisticated falsificationism', I'll have to remember that.
"So Joan, would you say that the results of your 100ks on the HST falsified Thornhill's 'electric universe' idea?"
"Yep Nereid, the electric cosmos people took the bull by the horns this time, and they posted some quantitative predictions on their website. We've crunched the results, and Scott and Thornhill were out by 245677 sigma!"
"That's cool Joan. Oh, by the way, did you dogmatically falsify them, methodologically falsify them, sophisticatedly falsify them, or did you develop some new falsification flavour?" :tongue2: )

But as Hugo (and many others) make clear, 'falsificationism' doesn't really tell you what scientist do :eek: (Hugo: "In summary, then, falsificationism in its various forms is an interesting idea but insufficient either to characterise science or solve the demarcation problem."). It is in this sense that I meant Popper has been 'discredited'.

For the avoidance of doubt, of course there is a real sociological and psychological aspect to the 'doing' of science - the people who 'do' science have human loves and hates; they crave money, sex, power, respect, etc to varying degrees; they had childhood traumas which were milestones in the development of the ids, egos, superegos and whatever; etc (you get the idea).To the extent that we can take a 'filter' from the UBVRIJHK filter set stored in the box next to the Nasmyth focus of Yepun that removes 'sociology' and 'psychology' lines from the incoming beam, let's do so.
 
  • #17
  • #18
Andre said:
No, he had a Ph.D in Physics But it appears that he knew what he was talking about.

Does it? The geologists say that the transition to plate techtonics roughly matches his schema of paradigm shift, but most other scientists assert the history of their own fields doesn't confirm it.
 
  • #19
Andre said:
No, he had a Ph.D in Physics But it appears that he knew what he was talking about.

Nope.

He only got a PH.D in history of science.
 
  • #20
Okay then Wikipedia should be corrected.
 
  • #21
According to an article I read a long time ago, Kuhn said he was in grad school studying physics when he became convinced that Aristotle's physics was not just stupid and wrong but represented a valid alternative mind set. So he switched his field of concetration to history of science.
 
  • #22
Hmmm [tossing pennies into the well]. Is truth not what is left after the impossible has been eliminated? [a Sherlock Holmes thing]. It is true scientists sometimes make 'leaps of faith' by removing the highly improbable from the field of play [after all, nothing is 'impossible']. But, I don't see what is wrong with that, despite the Popper objection. We can always revisit the improbable once the other options become even more improbable.
 
  • #23
Chronos said:
Hmmm [tossing pennies into the well]. Is truth not what is left after the impossible has been eliminated?

Possiblility is what remains after the impossible has been eliminated.
 
  • #24
The problem is to distinguish the impossible. Perhaps that formally mainstream declared impossible ideas may not be that impossible indeed. The only impossible thing here, may be getting that into the mainstream, or even getting it into the attention of the mainstream. Especially considering the human factors like "that-can’t-be-true" or "not-invented-here" or "trespassing-crackpots-will-be-shot".
 
Last edited:
  • #25
What happens when the "'falsification" itself is false? When the discovery of "cold fusion" was supposedly falsified by experiments, researchers continued to pursue their research anyway because they recognized that the experiments that supposedly falsified the theory were mistaken.

Another false "falsification" unfortunately was accepted and has hampered physics for over a century.

The famous Michelson-Morley experiment supposedly disproved the existence of an aether for light to be propagated through. Unfortunately the experiment suffered from a major logical flaw. Michelson falsely assumed that the movement of an energy pulse, or wave, through a medium was the same as someone swimming in water. He suggested that there would be some type of aether wind that would have the same type affect on the speed of light that the current in water had on a swimmer.

He ignored the fact that movement of a solid through a liquid required that the human body physically displace the water in order to move. A wave doesn't displace anything it just transfers energy from one element of the medium to another. The speed of a human body moving through water varies with the physical condition and swimming ability of the individual body. The speed of light is constant regardless of the physical qualities of the source of emission.

I recently did a search on Google and failed to find an equation for the speed of sound that included a variable for wind speed -- which would be a factor according to Michelson.

I'm not arguing here for the existence of an aether. That would be more appropriate in the physics forum on light. The point is that the Michelson-Morley experiment didn't disprove its existence.

Falsification of theories needs to be viewed with the same skepticism as "proof" of the theories. The experimentor might have overlooked some detail or might have proved a theory about the theory is wrong.
 

1. What is falsifiability?

Falsifiability is a concept in science that refers to the ability of a hypothesis or theory to be proven false through empirical evidence. In other words, for a hypothesis to be considered valid, it must be possible to design an experiment or observation that could potentially disprove it.

2. How is falsifiability used in science?

Falsifiability is a key aspect of the scientific method, which is the process scientists use to test and refine their theories and explanations. By requiring that a hypothesis be falsifiable, scientists can ensure that their conclusions are based on solid evidence and not just subjective opinions or biases.

3. What is the role of falsifiability in determining scientific credibility?

Falsifiability is an essential criterion for determining the credibility of a scientific theory or idea. If a hypothesis is not falsifiable, it cannot be tested or proven false, which means it cannot be considered a valid scientific explanation.

4. Has falsifiability been fully discredited?

No, falsifiability has not been fully discredited. While there have been criticisms of the concept, it remains an important and widely accepted principle in the scientific community. Some argue that falsifiability is limited in its application and that other factors, such as simplicity and coherence, should also be considered in evaluating scientific theories.

5. Can falsifiability be applied to all scientific fields?

Falsifiability can be applied to most scientific fields, but it may not be appropriate or relevant in all cases. For example, some areas of science, such as historical or observational sciences, may not lend themselves well to falsification. Additionally, some scientific fields may require different criteria for evaluating the credibility of a theory. Overall, falsifiability is a useful tool in most scientific investigations, but it is not a one-size-fits-all approach.

Similar threads

Replies
26
Views
1K
Replies
14
Views
847
Replies
47
Views
4K
Replies
10
Views
1K
  • Art, Music, History, and Linguistics
Replies
4
Views
957
  • Beyond the Standard Models
4
Replies
105
Views
10K
Replies
19
Views
6K
  • General Discussion
Replies
1
Views
2K
  • General Discussion
Replies
14
Views
2K
  • Beyond the Standard Models
Replies
11
Views
2K
Back
Top