[Edited to include my clarifications, as given in the Q&A below]
This question is in the realm of the "demarcation problem" of natural sciences, often called "the sciences", which includes physics. Mathematics is a formal science and, like philosophy, not natural science. However, we hope that the answer we seek here will be informative to all science disciplines, including mathematics and philosophy.
We are approaching this question in a different way, learning from Internet technical evolution and hoping to arrive at a better answer. We are using a methodology that is similar to that used to develop Internet technical standards, taking into account everything that works and also tentative implementations, and looking for consensus. Our first requirement to develop this "standard" (a "new sufficiency criterion", defining what does it mean for a theory or experiment to be scientific) , should be to promote interoperation; i.e., we want the standard to work not only in all branches of natural science but also in their dialogue, the interdisciplinary areas. What the standard considers to be scientific in biology should also be scientific in physics, in biophysics, and any other natural science area.
We already know that making clear (both new and known) predictions is useful as part of the "new sufficiency criterion" but this does not work as the entire "sufficiency criterion." We need more, and it was suggested we should also stress the "new" part -- that every theory needs to make predictions and solve problems that the previous ones could not do.
We already know that one cannot add consistency (even locally, within its own domain) to reinforce the "sufficiency criterion" because it already is an antecedent. A theory must be consistent within its own domain, before it can be considered a theory.
Karl Popper's answer, the falsifiability criterion, was that a theory is scientific if and only if it makes clear predictions that can be unambiguously falsified.
However, suppose that using Popper's falsifiability criterion in physics and natural sciences, can:
a. produce false positives (calling "scientific" a theory that is not); and
b. produce false negatives (calling "non-scientific" a theory that is).
If (a) and (b) are true, as I suggest they are, then Popper's criterion may just be "too blunt" to reliably define what is scientific or not.
Another problem with Popper's view of how science works, which also adds to the "bluntness" argument above, is that Popper's "logic of scientific discovery" falls apart in practice; scientific theories are not discovered as Popper (who never made such a discovery) imagines -- scientific theories have been and can be inductively inferred from experience.
Popper's "logic of scientific discovery" is disemboweled in the history of physics, with many textbook examples showing that experience can determine theory (i.e., we can indeed argue or infer from observation to theory, as Newton discovered the gravitation law, as Max Planck found the quantum).
In the posted question, please also consider modern natural sciences, including physics, quantum information, quantum cosmology, and biology. In a new area of natural science, what does it mean for a theory or experiment to be scientific?
I am hoping that your experience and reasoning can help in discussing these aspects of the "demarcation problem", including what is scientific, and how does scientific discovery work?
My own interest (and bet) in this question should be clear from the arguments above. I see a need to (1) recall Popper's falsifiability criterion; and (2) reject Popper's "logic of scientific discovery".
To be discussed in a forthcoming paper, my suggestion is that a new "sufficiency criterion" should be introduced, including what is already known, excluding Popper's falsifiability criterion, and introducing what I tentatively call "refutability" (the undeniable possibility that a statement is false). This is different from Popper's falsifiability criterion, although the word "refutability" is sometimes used as a lay synonym for Popper's "falsifiability".
Hi Edgard,
This is an interesting question. You may like to check a similar thread started by me on pseudoscience on my RG page.
Regards,
Rajat
Hello Rajat,
I checked the thread, thank you for calling my attention to it. Making clear predictions does not work as the sufficient criterion we need to see satisfied in order to qualify a theory as scientific. In this regard, Popper was correct in not accepting it either.
Cheers,
Ed Gerck
Hello George,
The posted question and the use of Popper's criterion to define what is "scientific" are both in the realm of natural sciences, often called "the sciences". Mathematics is a formal science and, like philosophy, not natural science. The contextual difference is that by "scientific" we traditionally restrict the subject areas to natural sciences only, which includes physics.
In physics, for example, at the end of the day it does not matter if the argument sounds coherent and scientific, is elegant, is beautiful, is intuitive, shows simplicity, is proposed by someone with high academic standing, is supported by a famous institution, has stood for a long time without any correction, or is defended by a very large majority of researchers. What maters is what nature has to say, the evidence.
Cheers,
Ed Gerck
Prediction in Physics is a necessary condition and fundamental in every model, but not the only one. Another condition, very necessary, which complements it, is that it needs to agree with the rest of models on such phenomenology. For instance, one cannot explain the disorder using the measurement of the entropy of the system if it is not assumed a closed system. This must be true for electrodynamics, quantum mechanics, relativity etc...
At the end, what is important, is that it has to be an economy of though in science, in such a form that every formula has to summarize many different experiments always in agreement with the rest of physical laws.
Daniel,
Thank you for your reply. First, there is no logical or natural law requirement that a theory that makes a prediction in physics must necessarily also agree with the rest (or even at least one) of the models used in that phenomenology. It may also create its own model ab initio.
Further, looking at various examples in physics, it does not follow that every formula has to summarize many different experiments always in agreement with the rest of physical laws, not even locally (ie, within its own phenomenology). You may see this as an ontology but not as a hierarchy (the evidence does not fit in a hierarchical model).
Thus, the conditions that you suggest are not necessary for a theory to be considered scientific. They are not even necessary for the theory's existence as a theory (scientific or not).
To be more precise, for every model in physics, such as special relativity, there can be many domains of research. In each domain, a theory cannot be inconsistent within its own domain, although the theory may be incomplete in that domain (e.g., there are unexplained phenomena). At any time, we may define a new domain where a theory is consistent within the domain, without regard for any other theory or domains using the same model. Thus, to be a theory, the theory must be consistent but the consistency domain does not have to be congruent with any other theory.
Although making clear predictions does not work as the sufficient criterion we need to see satisfied in order to qualify a theory as scientific, we cannot add consistency to reinforce the "sufficiency criterion" because it already is an antecedent.
Cheers,
Ed Gerck
Ed,
Let me try to respond your written:
“First, there is no logical or natural law requirement that a theory that makes a prediction in physics must necessarily also agree with the rest (or even at least one) of the models used in that phenomenology. It may also create its own model ab initio
What does it mean for a theory or experiment to be scientific? - ResearchGate. Available from: https://www.researchgate.net/post/What_does_it_mean_for_a_theory_or_experiment_to_be_scientific#view=567f01e37c1920a14d8b457a [accessed Dec 27, 2015].”
Answer:
The theory of gravitation of Newton contains Kepler's law. The electromagnetism of Maxwell has the previous laws of Coulomb, Faraday and Ampere. Newton`s Gravitation belongs to General Relativity as an special case, and so on.
Could you tell me one only independent theory in Physics?. The theories born because they are necessary for explaining new phenomenology and no as an artistic creation.
“Thus, to be a a theory, the theory must be consistent but the consistency domain does not have to be congruent with any other theory”.
Answer:
The basic concepts in every physical theory are the same in every theory that I know: energy, linear momentum, angular momentum, etc... From the quark to the graviton (if it could be found or not).
Daniel,
Thank you for your two comments. Let me answer your second one first.
I wrote:
“Thus, to be a theory, the theory must be consistent but the consistency domain does not have to be congruent with any other theory”.
In the first part of the quote, I am saying that a theory must be consistent in its consistency domain. I think we both agree with this statement. I also used this statement to support the central point of that reply to you, that consistency cannot be used to decide if a theory is scientific or not, as consistency is an antecedent.
In the second part of the quote, I am saying that a theory's consistency domain (eg, particles with negative mass) does not have to be congruent with the consistency domain (eg, mass is a positive number) of any other theory.
This is just pure logic of disjoint sets, which rounds up the quote.
Let me now address your first comment.
I wrote:
“First, there is no logical or natural law requirement that a theory that makes a prediction in physics must necessarily also agree with the rest (or even at least one) of the models used in that phenomenology. It may also create its own model ab initio."
In this quote, I wanted to introduce the idea that I eventually summarized at the end as “Thus, to be a theory, the theory must be consistent but the consistency domain does not have to be congruent with any other theory”, which I analyzed above.
Therefore, the analysis for this quote follows the same line of thought that I presented for the quote above (and that is why I presented it first).
The last phrase "It [a theory] may also create its own model ab initio" is an expression of the observation that physics models can be created at will (ab initio), that there is no need for congruency with other models, or with known reality.
Ab initio does not mean independent but "starting from the beginning". However, it does not need to start from "Bereshit bara" or "Let there be light..."
For an example in physics, how would an "Ab initio calculation of the neutron-proton mass difference" start? It can start as the eponymous article by S Borsanyi, published in Science 27 March 2015: Vol. 347 no. 6229 pp. 1452-1455.
You ask, "Could you tell me one only independent theory in Physics?"
I do not see this statement in my text. You are probably indirectly referring to the "ab initio" qualifier, discussed above, although ab initio does not mean independent.
About independence, we can say that any theory in physics has some degree of independence from other theories, otherwise it would not be a theory. Most physicists are also willing to accept the existence of things that are so independent that they cannot be proven.
About dependence, we can include energy (as we may wish to define it) and conservation of energy (if we want), and anything else. The good thing about definitions and theories in physics is that there are so many to choose from. But, certainly, we cannot choose all of them even if we restrict "all" to the set of theories not yet invalidated. That would not lead to a consistent result. Theories in physics cannot be fully dependent on all other theories, even valid ones.
Last, although I do not see any reference to my text, you wrote the conjecture that "theories [are] born because they are necessary for explaining new phenomenology and no as an artistic creation." I see how one could say that, immersed as we are in one form of sharashka or another, however A. Einstein wrote, "I am enough of an artist to draw freely upon my imagination."
Cheers,
Ed Gerck
Ed,
I am afraid that we can lost just in language. Let me to put you my opinion (ab initio) about your question again summarized in some points.
1. Every theory in Physics needs to contain the previous theories on the same phenomenon as special cases, e.g.
a) the relativistic mechanic transforms in classical mechanics when the velocities involved are much lower than the velocity of the light.
b) Einstein's gravitation transforms in Newton's one when the metric can be considered purely Euclidean and with time as a parameter.
c) the number of theories in Physics are really very few and in fact one of the last objectives is to find one of everything.
2. Every theory needs to make predictions and solve problems that the previous ones couldn't do. This allows to make economy of principles and ideas necessary.
3. Mathematics is one science absolutely necessary and without is impossible to understand the present Physics. Notice that Newton or Dirac made fundamental contributions to them as if they were a branch more of Physics.
Daniel,
Thank you for your reply. I see that all three points you mention can help answer or at least better qualify the thread's question, which is "What does it mean for a theory or experiment to be scientific?"
Let me preface my answer.
We are approaching this question in a different way, learning from Internet technical evolution and hoping to arrive at a better answer. We are using a methodology that is similar to that used to develop Internet technical standards, taking into account everything that works and also tentative implementations, and looking for consensus. Our first requirement to develop this "standard" (defining what does it mean for a theory or experiment to be scientific) , should be to promote interoperation; i.e., we want the standard to work not only in all branches of natural science but also in their dialogue, the interdisciplinary areas. What the standard considers to be scientific in biology should also be scientific in physics, in biophysics, and any other natural science area.
On your point #1. You and I, and most physicists, find it satisfying when a theory in physics contains the previous theories on the same phenomenon, even if just as a limiting case. However, we are willing to accept that when dealing with energy, the General Relativity does not contain Thermodynamics, the most fundamental theory of energy in physics, and starkly contradicts it; the aspect that General Relativity can transform into Newton's gravitation equation does not eliminate the contradiction with energy. Also, it is not commonly accepted in physics that we will ever find a Theory-of-Everything that contains all the previous theories. Therefore, if in the "standard" we require that "Every theory in Physics needs to contain the previous theories on the same phenomenon as special cases," then we would already start with exceptions. Also, considering interoperation, it may well be possible that a future theory in physics starts from a branch that we do not even know today. In the "standard" for what is scientific, just as we did when creating Internet technical protocols, we need to consider that in addition to things that we know we ignore, there are things that we ignore that we ignore, and plan for them.
On your point #2. Yes, every theory needs to make predictions and solve problems that the previous ones couldn't do. Although making clear, new predictions does not work as the complete "standard" we need to see satisfied in order to qualify a theory as scientific, it is part of the answer.
On your point #3. Yes, Newton, Dirac and many other physicists made fundamental contributions to mathematics, as if it were a branch of physics. I also agree that mathematics is the one science outside of natural science that is necessary, sine qua non, to understand physics. However, should we require the use of mathematics to be sine qua non for a theory to be scientific ? How about a theory that requires the use of mathematics that has not been invented yet? Again, In the "standard" for what is scientific, we need to consider that there are things that we ignore that we ignore, and plan for them. I suggest that the "standard" should make no demands on the methods used, just on the results.
Although the standard should also make no demands on mathematics, philosophy, and other areas outside of natural science, the notions developed in the standard should be applicable to them to some extent, further helping to promote interoperation (our first requirement). Linguistically and historically, we expect that all areas of science, whether based on phenomena or not, share the same abstract idea -- science -- and we hope to either capture that idea in the "standard" or, at least, not disavow it.
Cheers,
Ed Gerck
Ed,
I do not understand your sentence:
" the aspect that General Relativity can transform into Newton's gravitation equation does not eliminate the contradiction with energy". I do not know where is the contradiction with energy.
Anyway, let me try to summarize my point of view of your general question, trying to follow my experience in Physics or Mathematics:
A theory is scientific if can be reproduced independently of the observer and can explain many different phenomena in such a form that can predict results in every circunstance. Notice that here is contained mathematics ( for example Pithagoras theorem) or Physics.
Ed,
I do not understand your sentence:
" the aspect that General Relativity can transform into Newton's gravitation equation does not eliminate the contradiction with energy". I do not know where is the contradiction with energy.
Anyway, let me try to summarize my point of view of your general question, trying to follow my experience in Physics or Mathematics:
A theory is scientific if can be reproduced independently of the observer and can explain many different phenomena in such a form that can predict results in every circunstance. Notice that here is contained mathematics ( for example Pithagoras theorem) or Physics.
Daniel,
In GR, spacetime can give energy to matter, or absorb it from matter, so that the total energy is not conserved.This is a major contradiction with the fundamental theory of energy in physics, thermodynamics. Also, think about a red-shifted photon coming from another galaxy to us -- on arrival, the photon energy is hf and that was reduced because f was reduced compared with the starting f. Where did the missing photon energy (causing or as a result of the red-shift) go? There is no mass to absorb it.
I like your definition but I am concerned about the words "many", "observer", and "every". In particular, the word "every" tends to invalidate any sentence using it. We both agree on the the gist here; there may just be divergent views on what words to attach to the criterion. The requirement of repeatability (ability to reproduce) is necessary for a scientific experiment, but not for theory. Perhaps we need to divide the answer in two parts, theory and experiment.
In both cases, however, the result must be ostensibly refutable. No theory or experiment should assert that "this is the last word to ever appear on this matter". On the contrary, and I invite you to see the answers in my related question (link below). In science, we need to see a YES as NOT YET FALSE and a NO as COULD BE TRUE. When that is not understood, nature revolts and asserts itself. We can also find cases of this "nature revolt" in the formal science of mathematics, for example with the 1950's changed definition of a prime number. As part of a new criterion for "what is scientific", it seems to me that including refutability may also help reduce bias. People will tend to see things more in terms of "could be" than "is", a positive change for problem-solving as studied in social psychology (cf. Ellen Langer).
Cheers,
Ed Gerck
https://www.researchgate.net/post/Natures_fundamental_eg_fine-structure_and_cosmological_constants_have_values_in_the_small_highly_improbable_range_that_allows_life_to_exist_Why?_tpcectx=profile_highlights
Ed,
Sorry but you are wrong: space-time cannot give energy to matter at all. In general relativity Einstein tensor (geometry) is given by the energy-momentum tensor and not vice verse.
One different problem, and I thought that you refer to it, is that how is possible to define the energy-momentum tensor in a curved space-time because it is a Noether current associated to translations.
Daniel,
It is possible that I am wrong on the view that space-time can give energy to matter, or absorb it from matter, but I am in good company. You did not say what is your reply to the photon red-shift energy loss (see my previous answer) but it is possible you think (also in good company) that energy is conserved just fine when one includes the energy of the gravitational field along with the energy of matter and radiation. However, unless you find a way to define the density of gravitational energy, so that we can talk about a uniquely defined energy value at every point in space, then all one can talk about is the energy of the whole universe at once (which is not helpful to answer a photon's red-shift energy loss).
In summary, I think that although energy is conserved locally it is not conserved globally in GR.
The lack of global energy conservation in GR is, as I noted, a stark contradiction with thermodynamics, showing that theories in physics are not this "smooth tapestry" that you advance.
In my view, current physics theories cover the abstract space of all possible theories more like disconnected manifolds (a mathematical structure), and less like a "smooth tapestry", a connected surface where a path in one patch of that surface nicely continues to another patch. According to this "manifold model", global theories and global conservation laws may eventually be given up an replaced with local theories and local conservation laws. There could be no ToE, or one could have a ToE as a topological meta-system that projects into a (infinite?) set of disconnected manifolds.
I think one can take up the "manifold model" also in terms of Noether's theorem. and that is one of my pet projects, Perhaps, one could then finally prove or disprove the view that space-time can give energy to matter, or absorb it from matter. Until then, I am possibly not wrong on that.
Cheers,
Ed Gerck
Ed,
to doubt of the global conservation of the energy in the Universe belongs to the world of the entelechies: it is impossible to measure or to test in any form. Perhaps you are thinking in the topological properties of differential manifolds (Lie groups associated to the gauge symmetries?) as it happens within the horizons of a black hole. Although that is a great speculation perhaps in such a case we could speak about no global energy conservation in the space-time, but that matter needs a lot explanations out of this discussion. On the other hand I am not specialist at all in this subject.
Daniel,
Thank you. The point of our conversation, given by the posted question, is not whether I doubt the global conservation of energy. The point is "What does it mean for a theory or experiment to be scientific?"
In this sense, it is indeed useful to conceptually consider testing areas that (we think) may be impossible to measure or test in any form. Are these areas to be considered out of the realm of physics?
One answer to this question is that, whether or not we can observe something directly, contemplating its possible existence may allow us to understand how it might play a role in how the world works.
Another answer is that theories must be subject to potential reasoning and observational evidence that could persuasively decide in time whether the theory is wrong and lead one to abandon it as a scientific theory.
It may be possible to combine both answers, where a sound theory, with a clear domain of application and clear predictions (past and future, all valid in the past), and that
1. is NOT in disagreement with accepted theories, may be accepted as scientific even if we cannot at this time be able to think how to measure or test it in any form, and subject it to potential reasoning and observational evidence that could persuasively decide in time whether the theory is wrong and lead one to abandon it as a scientific theory; OR
2. is in disagreement with accepted theories, then it may NOT be accepted as scientific until such time that we may be able to think how to measure or test it in any form, and subject it to reasoning and observational evidence that could persuasively decide whether the theory may be considered a scientific theory.
Your thoughts?
Cheers,
Ed Gerck
Ed,
There is a third possibility that you do not consider: global conservation of energy in the Universe is always true by definition.
The Universe doesn't allow to scape anything from it (included the energy). Thus is tautology global conservation in the Universe of everything that you want. Thet's just definition.
Hsppy, Fruitful and Healthy New Year 2016!
George,
Used to be. You will find in some old tables that 1 is listed as a prime number.. Circa 1950, some mathematicians got tired of writing "any prime number except 1" in theorems and decided, one by one in DIY fashion, that 1 would no longer be a prime number — although 1 satisfied all the conditions to be a prime number. There are many other examples in pure mathematics, it is also an experimental science.
Happy 2016!
Cheers,
Ed Gerck
Daniel,
You wrote, "The Universe doesn't allow to scape anything from it (included the energy). Thus is tautology global conservation in the Universe of everything that you want. Thet's just definition."
A definition that cannot be verified by physical experiments? That's what we are discussing here -- should it be expelled a priori as metaphysical mambo-jumbo or can it be properly used in a scientific context?
If one defines that the Universe is a closed system, and yet energy is disappearing when a photon red-shifts coming from another galaxy, in physics one still does not need to change the definition although one must still be open to that possibility! All that one needs to do in physics is to provide a verifiable, refutable explanation for the energy loss or energy balance in photon red-shifting.
If the definition cannot be changed (refuted) then it is just not scientific, but it may be mathematical, or it may be metaphysical mambo-jumbo. That is not necessarily bad or irrelevant to physics. For example, we know that pi does not physically have the value 3.141592... that mathematics says that it has, for more than one reason. It is not scientific, it is quite arbitrary, yet physics uses it.
Cheers,
Ed Gerck
Ed,
I think that we agree now, to speak about the global conservation of energy in the Universe is not a scientific issue at all, as I tried to argue you some posts before . It is assumed its conservation and without possibility of testing it (ok, this is not even science).
The problem, if we can speak about this question as a problem, is if the energy can be defined globally in the Universe or what is the same, if the symmetry conditions of the time allow it. But when we descend to theses details the important is its local conservation.
Ed,
There is another subtlety related with the term "global". This can refers to the non trivial topology of the space-time and this is related with the singularities associated to the metrics.In general, a manifold cannot be covered by a single coordinate system.An atlas of coordinate systems is required, and it reflects the global prop-erties of the manifold. An individual coordinate system can be limited by its singularities, that may be seen as singularities in the metric topology with respect to the coordinate-based topology. For instance the Schwarzschild metric has two singularities en r=0 and r=2m being m the mass introduced in the metric solution of the Einstein's equations.
I suppose that you do not want to enter in these problems where the concept of local and global is quite difficult to distinguish, at least for me, that I am not a specialist in General Relativity and Gravitation.
Daniel,
Thank you for your two replies above, that I agree with and am commenting together. The terms "local" and "global" may be more of a help to our imagination to understand nature than a feature of nature. If this is true, then the distinction of local vs. global is not fundamental and should change with theory, observer, experiment, and interpretation. I think we can see that in quantum entanglement and GR, just to name two areas.
You perceive correctly that I am hesitant to enter into these topics at least here in this Q&A. Mainly, that is because while they have been useful for us to discuss the boundary of "what is scientific" (the posted question), they should not lead us off topic. Perhaps in a new Q&A?
I think we also agree that nature shows an extraordinary flexibility in how things are accomplished, very often surprising our imagination and what we considered experimental facts. Pascal said that imagination tires before nature. I see that it is timely for natural science to be recognized in a far wider domain -- as we see in practice -- than Popper (with no experience of note in natural science), Cartwright (ditto), and other philosophers wanted to define as "scientific".
The question (here) is what this wider domain should be so that physics does not become pseudo-science, mathematical conjectures, or a mathematical-philosophical mambo-jumbo. This question is at the frontier of natural science today (eg, see http://www.nature.com/news/scientific-method-defend-the-integrity-of-physics-1.16535). I suggest that, at least for the time being, we in physics are better qualified to provide an answer, an answer that natural science can use, than any other natural science branch or philosophy. I also suggest that it is high time for philosophy to look at the Flynn effect (psychology), algebra, quantum mechanics, and the Internet, and realize just from these four examples that we can think things out much better and faster, individually and collectively, than the ancient Greeks could, God bless their souls.
Cheers,
Ed Gerck
Are scientific theories discovered or observed? I think sometimes it is actually discovered. Perhaps in the pure mathematics area, or even philosophy... Then at some point, someone says... Hey... Maybe that pattern actually applies to this thing we have in the real world? They may have reason behind it... I think of Boltzmann's canonical ensembles which I would think is almost surprising that it leads to thermodynamics. Or perhaps the discoveries of Euler, and the use of exponents of imaginary numbers, which were found to be descriptions of rotational motion, and hyperbolic rotation, for sine, cosine, sinh, and cosh. Sometimes a theory comes about through direct observation, but sometimes, the phenomenon may be known and the math may be known separately for some time before anyone thinks to link the explanation with the pattern in nature.
Hello : Yes, I take the view that all laws of nature are discovered, and none are invented. The universe is billions of years old, the imagination is human and very, very, very recent.
But scientific theories are not discovered as Popper (who never made such a discovery) imagines -- scientific theories have been and can be inductively inferred from experience. That experience may include direct or indirect proofs, may be totally mathematical in representing the data as a model.
- Badges
- Science topic
- Similar topics
- Physics