Talk:Falsifiability/Archive 5

Archive 1

←

Archive 3

Archive 4

Archive 5

Archive 6

Archive 7

→

Archive 9

Statement of the obvious

We are told that the concept of falsifiability was "introduced" by Popper. It is a statement of the obvious. Everyone has always objected to a theory that cannot be disproved.— Preceding unsigned comment added by 86.146.216.125 (talk) 12:54, 27 May 2019 (UTC)
In 1966, Popper noted that he wished he had known about Peirce's work earlier. -- 13:21, 27 May 2019‎ 86.146.216.125

86.146.216.125 -- Popper didn't really invent the idea (though he explored it more systematically), but no, it's not entirely obvious. Many scientists would have said that the best theory is the one that most closely fits experimental observations in the most parsimonious way... AnonMoos (talk) 13:47, 27 May 2019 (UTC)

The usual logical notion of falsifiability which is the capacity to be falsified is as old as the notion of falsification, which was known in ancient Greek and Indian philosophies. However, the epistemological definition of the term falsifiability was given by Popper. This is well known. Charles Sanders Peirce coined the term "fallibilism" in the late nineteenth century and these are related concepts. However, these terms were coined in completely different contexts. The article is about the epistemological notion of falsibiability, as introduced by Popper, but we can certainly mention how it relates to Peirce's view, etc., as long as we have valid sources. Dominic Mayers (talk) 14:39, 27 May 2019 (UTC)

Perhaps the fact that falsification (and thus falsifiability) was known in the ancient times may appear beside the point, because it was not contested here that Popper redefined the term falsifiability, so that it is something different. The issue is whether it is new with respect to Pierce, etc. Indeed, simply because Popper defined or redefined the term falsifiability does not mean by itself that he introduced a new idea. Certainly, new knowledge is always related to previous knowledge and nothing is never totally new, but some contributions are considered more important, because at the time they changed the way of thinking. These contributions deserve separate articles in Wikipedia. How do we know that Popper's falsifiability is such a contribution? We know it because so many books and articles, secondary sources, acknowledge this contribution. For example, I have a book in front of me (and there are many others that do the same), which has a chapter written by W.V. Quine. The chapter is entitled "On Popper's negative methodology". This illustrates that many illustrious mathematicians and philosophers of science have attributed falsifiability and its associated methodology to Popper, even though they obviously knew that it was not entirely new. Dominic Mayers (talk) 15:33, 27 May 2019 (UTC)

Re-write needed

Respectfully, a subject expert needs to reorganize the text to eliminate run-on sentences and grab bag paragraphs. Incrementally added text makes the article hard to understand.

The article sounds like a college term paper being padded for length. For example, the names of Karl Popper and other authorities are repeatedly inserted into the text unnecessarily. Please use footnotes to refer to authorities.

Also, some paragraphs now contain multiple ideas, which overcomplicate the text. Please break paragraphs into groups of ideas.

The issue is not accuracy. The goal is organization and clarity. Tdk408 (talk) 04:26, 28 December 2019 (UTC)

The awkward tag

The awkward tag out of context at the start of the article is confusing. Such a tag makes sense when used after a sentence such as explained in its usage: https://en.wiki.x.io/wiki/Template:Awkward . The first step if a reorganization is needed, is to propose changes in this discussion page, at the least point at specific sentences or paragraphs so that we have some context to discuss it. Dominic Mayers (talk) 03:05, 30 December 2019 (UTC)

I written my thoughts about why we need a plan to proceed Dominic Mayers (talk) 23:18, 2 January 2020 (UTC)

Clearer lead

The article currently starts:

A statement, hypothesis, or theory is falsifiable if its assertion contradicts data (more precisely, a basic statement) that is otherwise possible.

This wording is trying hard to be technically correct, but at the expense of being understandable by the ordinary reader, for several reasons:

• "its assertion" seems to be used to mean "whether it is true"
• it is unclear what it means for its assertion to contradict data
• mentioning the technicality of "basic statement" is premature here; for one thing, "basic statement" isn't defined
• "data" is unclear
• "that is otherwise possible" is unclear.

It also doesn't give enough context.

Let me propose something simpler:

In philosophy, a claim is falsifiable if there could in principle exist an observation which shows it to be false. Many theories in the philosophy of science use falsifiability as a crucial criterion separating science and non-science; making this precise has occupied many philosophers.

It would also be nice to have a vivid, concrete example somewhere in the lead, e.g.:

The claim "all ducks are white" is falsifiable—and indeed false—because observing one black duck contradicts it. The claim "all ducks weigh less than a ton" is falsifiable but has never been falsified, and is indeed surely true. "All ducks are mortal" cannot be falsified in principle—even though it is doubtless true—because there is no way of knowing whether any of the currently living ducks will ever die.

I'm sure that the definition and the examples above can be improved... but my point is simply that the article needs to start with a general, understandable statement. Later on, of course, the article can go into necessary details, technicalities, and caveats. After all, the lead can't possibly summarize all the work on the topic! --Macrakis (talk) 18:11, 9 January 2020 (UTC)

Small Note

By the way, falsifiability is NOT a criteria that directly separates science from non science. If you want to write that, then you are biased with an agenda, which might be honorable, but still this bias would prevent you from presenting falsifiability for what it is. Falsifiability only applies to theories, not to scientific methods that accept or reject these theories. What can be said is that a method that would accept unfalsifiable theories can hardly be scientific, but this is a meta criterion that uses falsifiability. It's not falsifiability. There is a nuance and it is important.

Dominic Mayers (talk) 22:58, 9 January 2020 (UTC)

Agreed. "Naive falsifiability" is problematic in all sorts of ways, and using it as a blunt instrument to separate science from non-science doesn't really work in any sort of straightforward way, as a whole series of philosophers have pointed out over the past 60+ years.

My point was that the lead to the article needs to present the basic idea of falsifiability in a simple, comprehensible way. The basic idea is along the lines of: if your theory can't be contradicted by observation, it's not saying much. I really do prefer the term "observation" to "data", because "data" can mean many things -- not just observational data, but also any information at all. A book of random numbers is data in the latter sense. Maybe it suffices to rename it "observational data".

The current wording is difficult to understand, even for someone (like me) who has actually read Popper, Kuhn, Lakatos, Feyerabend, et al. What does it mean for data to be "possible"? Why talk about the "assertion of a statement"?

The core idea of falsifiability is not that hard, and the introduction to the WP article should be understandable by an intelligent 10th grader (or 10th grade teacher!).

It also seems important to mention the role of falsifiability as a demarcation criterion in the lead. --Macrakis (talk) 23:26, 9 January 2020 (UTC)

My motivation for "data" was to refer to a common concept in practical science and data is such a concept. In particular, it includes the outcomes of complex processes and computations such as in magnetic resonance imaging, etc., but maybe you are right that it was too general. Observational data is better. We agree on this. I am thinking more about the other points that you made.

Dominic Mayers (talk) 05:58, 10 January 2020 (UTC)

Not much can be said based on the naive definition, only confusion.

The problem is that we naively think that it is sufficient to have a definition that we know how to use to demarcate falsifiable theories from others.

• All swans are white is falsifiable, because we can conceive the observation of a black swan (in fact they exist).
• All swans weigh less than one ton is falsifiable, because we can conceive the observation of a swan that weigh two tons.
• All swans are mortal is not falsifiable, because we cannot know if the swan is immortal. It's not an observation.

Simple definition: A statement is falsifiable if we can conceive an observation that contradicts it. That seems perfect and at the level of Wikipedia. This is wrong. The definition was not introduced to be used in this naive manner. This is a complete misconception. It was introduced to exhibit the logic of scientific discovery. In accordance with Popper, science is a complex process that use hypotheses (theories that are proposed) and refutations. The generation of hypotheses cannot be explained by logic, because induction is not logical. The only part that use logic is the refutations. The purpose of falsifiability is to show the purely logic content behind the refutations. The naive definition does not work for this purpose, because it does not has the key ingredient. The illusion that is created by the naive definition is misleading. Worst, this propaganda for the naive definition has contributed to many unfair criticisms toward falsifiability. So, if we give the naive the definition in the lead, we need to provide the proper context in the lead as well. I know, this definition is given in a few sources, but rarely out of context. Most sources avoid this naive definition.

What is missing? The link between the logical statements and the actual observations is missing. Given that the role of falsifiability is to exhibit the purely logical refutations that correspond to the actual refutations, this link is crucial. The whole discussion depends on it. This link requires the notion of basic statements, which are built using a background theory or a background knowledge. Popper discusses in details the issues of the empirical basis and what are the basic statements before he gives the definition.

Dominic Mayers (talk) 11:42, 16 January 2020 (UTC)

Unsourced debate about the subject of critical rationalism

This paragraph raises the question whether critical rationalism is about scientific theories only or about other things as well, but the only source provided is not convincing at all and I could not find any source that refer to Criticizability explicitly.

In contrast to Positivism, which held that statements are meaningless if they cannot be verified or falsified, Popper claimed that falsifiability is merely a special case of the more general notion of critical rationalism, even though he admitted that empirical refutation is one of the most effective methods by which theories can be criticized. Criticizability, in contrast to falsifiability, and thus rationality, may be comprehensive (i.e., have no logical limits), though this claim is controversial, even among proponents of Popper's philosophy and critical rationalism.

In the reference provided, Popper does not say that "falsifiability is merely a special case of the more general notion of critical rationalism". First, it's not a special case, because critical rationalism is about the method whereas falsifiability is about the logical structure of the theory. Second, it's not clear that when Popper wrote "theories or guesses" that he meant that we should apply the methods of critical rationalism to other things than theories.

Dominic Mayers (talk) 20:01, 17 January 2020 (UTC)

What I did to improve the section Problem of induction

I removed the long paragraph that used "naive falsificationism" in a confusing way and then only listed a sequence of scientific theories, as if it was a proof of something. I placed the criticisms in their proper context. As a bonus, we have the explanation for the correct non dogmatic definition of falsifiability. The most important point though is that this is now very easy to source. I will be happy to add any citation needed. Dominic Mayers (talk) 23:20, 20 January 2020 (UTC)

Restructure and edit

So I've re-written a few slabs, reducing length, removing waffle and reinstating a structure in the lead that is reflected in the article.Banno (talk) 23:20, 18 January 2020 (UTC)

Thank you. This is a good cleaning and a better organization. There are many things that you removed that I did not have the courage to remove. I intend to reinsert some important and well sourced points that you removed, but it's ok that you removed them, because we need to think about where they fit. My guess is that they can be reinserted in the current structure, perhaps with an extra section for the definition, but maybe not. We will see. Dominic Mayers (talk) 03:31, 19 January 2020 (UTC)

Cheers. I'd encourage the reintroduction of some of that material. My aim was not to remove it but to restructure. Banno (talk) 23:22, 22 January 2020 (UTC)

Confusing section: Demarcation problem

We have to rethink this section. We could rewrite the second paragraph as a critic of naive falsificationism. We would swap the first and third paragraphs, because the third paragraph leads to naive falsificationism, because it is about rejection, whereas the first paragraph explains the usefulness of the criterion in a way that works in sophisticated falsificationism, where we only assess a theory with falsifications, but the final decision to reject is more complex and depends on failed falsifications (corroborations) as well. The section will then be a description of the demarcation problem to introduce the criterion, which should be provided somewhere. Any other idea? Dominic Mayers (talk) 18:21, 21 January 2020 (UTC)

I don't see how a reader would understand demarcation before naive falsification; and that this should be in the problem of induction. Go ahead with your edits, so I can see where you are going. Banno (talk) 00:01, 23 January 2020 (UTC)

I will go ahead, but before I would like to better understand what you are saying here. Can you expand a bit and be more specific about what aspect, if any, of my edits in the article or ideas in the above paragraph you are referring to? Dominic Mayers (talk) 03:50, 23 January 2020 (UTC)

Sokal and Bricmont

I'm questioning noteworthiness; but when investigating, I found that the quote provided is wrong. Specifically, the sentence " ... [but] the history of science teaches us that scientific theories come to be accepted above all because of their successes." was not found in the book. Nor could I locate support for "They further argue that falsifiability cannot distinguish between astrology and astronomy, as both make technical predictions that are sometimes incorrect." in the book. Hence, my edits. Banno (talk) 20:34, 24 January 2020 (UTC)

Does Wikipedia need this article?

Is Falsifiability a thing?

It is apparent that the logic of falsification, posited as an answer to the problem of induction, is not presented at all clearly in thie article. Instead the push seems to be for an article describing falsification as the principle demarcation between science and non-science. That is, the article is now partisan.

Neither Stanford nor IEP have articles on falsifiability. The topic is subsumed into the articles on Popper and on Scientific Method.

What do folk think of moving the bulk of this material into a new article on Falsificationism, which would be an historical account of Popper, Lakatos, Feyerabend and the relevant controversies, with this article becoming a mere redirection? Banno (talk) 21:42, 16 January 2020 (UTC)

You raised two independent questions. Is the article biased regarding the problem of demarcation vs the problem of induction? Does falsifiability deserves an article in Wikipedia? These two questions are to a large degree unrelated. I will first answer the second question. The definition of falsifiability contains a lot of subtleties and it is considered the core of Popper's epistemology. It is a part of critical rationalism. My answer is that we should look for a better integration between falsifiability and critical rationalism. These two articles should support each other and it's not the case now. We cannot decide if falsifiability should be a separate article before we worked on how these two subjects can support each other. Falsifiability should be a tool used in critical rationalism and critical rationalism should provide an extension of falsifiability, a larger context.

Regarding the bias, it's not like there is an agenda being pushed here, not at all. Every author, I am sure, is interested in covering all facets of falsifiability. There is no issue here. Yes, maybe a facet could be covered more. Great, let's do it. It's not an issue. By the way, the original motivation for falsifiability was the demarcation of science vs non science. But it is true that the problem of induction is very much related to this problem, because the previous solution (in the Vienna Circle, Berlin Circle, etc.) to the problem of demarcation was based on an attempt to describe science as some form of induction (we start from the truth of observation statements and we build from there to have more complex laws, etc.). This is the way falsifiability addresses the problem of induction. It does not make induction possible. It does not solve the problem that way. It solves it because the context at the time was that some form induction was considered essential to separate science from non science and this lead us no where. Falsifiability solves the problem, because now, Popper says, it is not necessary. What do you expect as a solution to the problem of induction? What exactly you feel is missing?

Dominic Mayers (talk) 01:21, 17 January 2020 (UTC)

I'm not making a philosophical point, and agee with much of what you say. I'm asking wha thte best arrangement for an encyclopaedia might be. Given that other encyclopedias do not include an entry such as this, it seems to me that we ought at least take pause and think about it being included here. Banno (talk) 01:50, 17 January 2020 (UTC)

Perfect ! I like that. I asked the same question to myself some time ago. There is an advantage in merging it in critical rationalism: it is naturally introduced by the larger article. It would naturally be integrated with critical rationalism. Yes, but it remains that it is important enough to be a separate article. It is natural to expand on falsifiability in its own article. By the way, some encyclopedia of science covers falsifiability. I think I have two on my bookshelf that covers falsifiability. One of them has an article by Alain Boyer. So, on that respect, we are perfectly fine. But, sincerely, I like very much that you asked the question. My proposal is that we work toward a better integration with critical rationalism. This is a challenge that must first be addressed. It will make it easier to see if we should merge them and how we would do that, if we should. Dominic Mayers (talk) 02:21, 17 January 2020 (UTC)

Excellent. I guess we proceed by working out what belongs here, and what belongs in critical rationalism. I think this is the place for a discussion of the logic of falsifiability, and a place to contrast it with verificationism. Overall, I agree that this article ought be a brief jump-off point to a wider understanding of the topic. Banno (talk) 09:54, 25 January 2020 (UTC)

OK, I'm going to take this point as moot, and instead sugest working on the article in such a way that it would not be a concern if it's title were so changed. Banno (talk) 02:31, 18 January 2020 (UTC)

Unverifiable and unclear paragraph

The paragraph that starts with "In place of naive falsificationism, ..." is not sourced. The first sentence suggests that the previously mentioned issues are the defining characteristics of naive falsificationism. Who is the well known philosopher that used "naive falsificatism" in this manner? Certainly not Lakatos who has popularized the concept and first carefully analyzed it. The remainder of the paragraph "Popper envisioned science as ... " is essentially weak propaganda for Popper methodological and sophisticated falsificationism, because even a dogmatic falsificationist (see Lakatos 1978, p. 13-14) can envision the series of falsifications described in the paragraph. The overall section is problematic. It contains only one citation and it seems incorrectly used. Dominic Mayers (talk) 18:52, 20 January 2020 (UTC)

Naive Falsification was introduced as a structural element, long ago. The intent, at least on my part, was that the article show issues with the verifictionism of the Viena Circle, mention how verificationism seems to rely on confirming the consequent and explain how Popper's use of Modus Tolens overcomes this. I'd intended to lead on to a discussion of the stats of falsification. but eventually moved on without finishing. Character drift more than removed its usefulness.

I suspect the term is from SSR, but am not in a position to verify. It is certainly used.

I now wonder if it might be better to keep this article relatively brief; a jump of point to the more detailed content that exists elsewhere on Wikipedia. Banno (talk) 23:56, 22 January 2020 (UTC)

To my knowledge naive falsificationism is not used in SSR. The term is often attributed to Lakatos, even if Lakatos attributes it to Khun. Perhaps Lakatos meant that Khun only considered naive falsificationism. What is clear is that Lakatos clearly defined the term and the term is strongly attached to Lakatos. Dominic Mayers (talk) 21:01, 23 January 2020 (UTC)

You are correct, it is not in SSR but in Lakatos' repy to SSR. 15:43, 25 January 2020 (UTC)Banno (talk)

dogmatic falsificationist

Why aren't we using Lakatos' term naive falsification? After all, even a sophisticated falsificationist can be dogmatic.Banno (talk) 05:02, 27 January 2020 (UTC)

Important question. Thanks for asking. Lakatos defines both dogmatic falsificationism and naive falsificationism. They are different concepts. Both are important. Dogmatic falsificationism has nothing to do with the naive claim that if a theory is methodologically (i.e., not dogmatically) falsified, then it will be rejected from science. This naive claim is naive falsificationism. Dogmatic falsificationism is different. It is the dogmatic belief that we can have faithful (in my own terminology) observations of the reality, i.e., that if we write a basic statement that corresponds to an observation, then it's true. Lakatos and Popper explain that in reality, it requires some decisions taken by agreement among scientists. So, to sum up, we move out of dogmatic falsificationism to enter methodological falsificationism and we move out of naive methodological falsificationism (or simply naive falsificationism) to enter sophisticated falsificationism (or simply falsificationism). Dominic Mayers (talk) 07:22, 27 January 2020 (UTC)

Lead

Long discussion about the lead

See Wikipedia:Manual of Style/Lead section. Specifically: the lead should be reduced to a few short sentences; the lead should stand on its own as a concise overview of the article's topic; The lead should establish context; it should summarise the most important points, including any prominent controversies. So here are the key points it should contain... Falsifiability is an answer to both the problem of induction and the demarcation problem It was introduced and developed by Karl Popper It is now a common, if not uncontroversial definition of what is scientific an what is not As such it has had a prominent role in various controversies and legal cases. Thoughts? Elsewise I will edit as above. Banno (talk) 02:53, 18 January 2020 (UTC) I agree that these points should be included. I think you will raise controversy to say that it is uncontroversial. Do you intend to remove entirely the definition and only says these things? I would not object, but this a bold decision and some will be tempted to add the naive definition. What I would not agree with is to only have the naive definition in the lead, as if it contained the essential idea. Better not have the definition at all in the lead. I suggest that you add these points, but keep the definition together with the necessary explanation of what "can conceive an observation" means. We could remove the second example "All men are mortal". This will make it shorter. Dominic Mayers (talk) 03:52, 18 January 2020 (UTC) Most things interesting about falsifiability depends on its dependence on the background knowledge. One example is this statement by Thornton: "For Popper, however, to assert that a theory is unscientific, is not necessarily to hold that it is unenlightening, still less that it is meaningless, for it sometimes happens that a theory which is unscientific (because it is unfalsifiable) at a given time may become falsifiable, and thus scientific, with the development of technology, or with the further articulation and refinement of the theory. Further, even purely mythogenic explanations have performed a valuable function in the past in expediting our understanding of the nature of reality." — Thornton, https://plato.stanford.edu/archives/sum2017/entries/popper/ This makes no sense if falsibiality does not depend on the background knowledge, because it says that when the technology changes (thus the background knowledge changes), falsifiability changes. Dominic Mayers (talk) 05:50, 18 January 2020 (UTC) To sum up, I am not opposed to an article that first gives the naive definition and then later the correct definition, which refers to the background knowledge or to basic statements. My objection is only that the lead cannot just have the naive definition and give the impression that it is the definition. This would be misleading. Dominic Mayers (talk) 12:49, 18 January 2020 (UTC) To this: [1] I'd said that I would edit, but anything added reduces the clarity of the intro or is my own work. What needs to be said is something more verbose... Suppose we find a new creature, a zwan. The zwans we have seen have all been white. We can formulate the hypothesis that all swans are white. The opening sentence of the article is exactly true: some observation might show it to be false, but that observation has not yet been made, and hence the hypothesis is [falsifiable], but at this stage it is not falsified. So instead I will leave things as they are, and ask if anyone else shares (User talk:Dominic Mayers) misgivings.Banno (talk) 00:47, 24 January 2020 (UTC) I think you might confuse two things: reducing clarity and making clear an existing issue. When you make the issue apparent, people are getting confused for a good reason: they thought it was clear and simple and they realize that they cannot explain it in concrete terms. There is no other way to explain it than referring to the abstract notion of a language that allows to conceive events that are not real. It is also very obvious that it's the way Popper sees it, explains it and also insists that we do the same. He always insisted that we distinguish between the logical criterion, which is only at the abstract level of statements, and the methodological rules. Dominic Mayers (talk) 01:26, 24 January 2020 (UTC) Your attempt to explain it in terms of an observation that has not yet been made does not work. If the contradicting observation can possibly be made tomorrow, the law is false (because it is implicit in the concept of a law that, if it is not true tomorrow, then it is false.) One way that would work is to look for a contradiction by going back at a time where the law was completely ignored and then we ask what is possible in accordance with the background knowledge at the time. It works, but this background knowledge is the samething as the underlying logic + its physical interpretation. Dominic Mayers (talk) 01:45, 24 January 2020 (UTC) I've no clear understanding of what your point is. So, in order to proceed, what is your sugested replacement?Banno (talk) 11:13, 24 January 2020 (UTC) I edited a bit my answer above to make it easier to understand. There is a part of me that feels ok with A claim is falsifiable if some observation might show it to be false., but it's not logical. This feeling must come from the fact that the statement in itself is simple, even though it makes no sense as a criterion. There is another part of me that explains why this sentence does not work logically. I also remember my first experience with a similar sentence when I read for the first time this article many years ago. I was confused. It made no sense. You have tried to explain it with : some observation might show it to be false, but that observation has not yet been made,... Do you understand why this cannot be how falsifiability works? Think about the fact that falsifiability is for laws of science that do not change in time. If falsifiability meant that the law can be shown to be false tomorrow, then it would mean that the law is false now. It has to be non ambiguous that the contradicting event is purely formal. If the contradicting event is not formal, but real, no matter if it is only possible and not actually observed yet, the law is false. I understand how one might feel, because we refer to an abstract space of formal observations to define falsifiability. I understand the feeling. One might feel "we make it hard to understand, not clear", but at the same time, it's the only way that it makes sense. We must take into account that Popper always insisted that we separate the logical criterion from the methodological rules. Dominic Mayers (talk) 14:31, 24 January 2020 (UTC) Seems to me you are overthinking it; but I am content to go with "claim" since it also avoids a pending a/an edit war. Banno (talk) 18:25, 24 January 2020 (UTC) My point has nothing to do with claim vs hypothesis. It's the same concept whether we use claim, hypothesis, statement, theory, etc. My opinion is "theory" is the best, but I don't really care. I am not overthinking it. You tried to address the problem that I explained by referring to an observation that has not yet been made. You did not answer my question. Do you understand my explanation why this cannot work? Dominic Mayers (talk) 02:49, 25 January 2020 (UTC) I'm not interested in a philosophical debate; there are forums for that. Modality is not to do with time, but logical status. A statement that has been falsified is falsifiable. A statement that has not been falsified but will be, in the future, is falsifiable. A statement that never was, and never will be, falsified, might never the less be falsifiable. And if you do not understand and agree with that, you should not be editing the article. And I still have no idea if you have an objection to the opening sentence, yet alone what that objection might be. As I asked before, what woudl you like to see there? And if you do not have an alternative, then there is no point to this conversation. Banno (talk) 03:06, 25 January 2020 (UTC) Unfortunately, you have not written anything that shows that you understand the problem when we refer to a falsifying observation, whether it is in the future or in the past. You wrote A statement that never was, and never will be, falsified, might never the less be falsifiable., but it does not say the key point, which is that a statement that cannot in anyway be falsified (in the future or whatever) by an actual observation might nevertheless be falsifiable. You see the apparent non sense: we say that it cannot be falsified (which is more than only saying that it will never be falsified) and yet it might be falsifiable. That's the key point. The "falsified" corresponds to one meaning of falsifiable. The "falsifiable" corresponds to a second meaning, the technical meaning. My objection is still the same. The first sentence says that it is falsifiable if there might be an observation that shows it to be false. If we don't explain that the falsifying observation is taken from an abstract space of observational statements, this makes no sense. An intelligent 10th grader will be very confused by that. Dominic Mayers (talk) 03:28, 25 January 2020 (UTC) Also, this first sentence in the lead is not at all represented in the body of the article. The current article is about falsificationism and written in a way that avoid the definition of falsifiability. So, to respect the body, the lead should only be about falsificationism. Dominic Mayers (talk) 03:28, 25 January 2020 (UTC) We can work together toward a lead that makes sense. Why do say that it had to be me that write the solution? Dominic Mayers (talk) 03:28, 25 January 2020 (UTC) So, third time, what's your suggested alternative? Banno (talk) 04:30, 25 January 2020 (UTC) For anyone who does not see the first sentence in the lead as a serious issue, the abstract nature of the correct definition will not be acceptable. It would feel like "less clear", but only because we confuse "less clear" with "making apparent an issue that we think should not be raised". There is no rule that says that someone that mentions a valid issue must provide the solution. So, please do suggest a solution, if you understand the issue. If you don't understand the issue, then the role of a talk page is to discuss issues like that, which are important in the article. Dominic Mayers (talk) I've been watching this for a while without wanting to get involved, but I really think there's no serious problem with the current first sentence and your attempts to say what is wrong with it are much less clear than that sentence itself. Falsifiability is the ability to be shown to be false. Maybe "Falsifiability is the ability of a claim to be shown to be false."? Or "Falsifiability is the ability of a claim to be shown by observation to be false."? Or "Falsifiability is the ability in principle of a claim to be shown by observation to be false."? Or "Falsifiability is the ability in principle of a claim to be shown by observation to be false if it is false."? I feel like this is becoming an unwieldy sentence the more clarifying clauses I add to it. --Pfhorrest (talk) 07:31, 25 January 2020 (UTC) I know it's clear. It's clear, but it's not what Popper meant. It's non sense as a demarcation criteria. If your only criteria is clarity and it does not matter that it is wrong, then here is the problem. By the way, I agree that adding "in principle" or playing with words in a similar manner is not the way to go. It makes it worst in a way. Yes, what I explain is not as clear, because it's a concept that relies on a lot of background. It is to be expected that it subjectively appears less clear than the wrong naive but "clear" definition. Dominic Mayers (talk) 07:45, 25 January 2020 (UTC) It is what Popper meant, if not all of what Popper meant. You can't put the whole article in a singe sentence. If you don't have an alternative, your opinion becomes irrelevant. Banno (talk) 09:59, 25 January 2020 (UTC) If it is what Popper meant, explain this quote from Popper: "We must distinguish two meanings of the expressions falsifiable and falsifiability: "1) Falsifiable as a logical-technical term, in the sense of the demarcation criterion of falsifiability. This purely logical concept — falsifiable in principle, one might say — rests on a logical relation between the theory in question and the class of basic statements (or the potential falsifiers described by them). "2) Falsifiable in the sense that the theory in question can definitively or conclusively or demonstrably be falsified ("demonstrably falsifiable"). "I have always stressed that even a theory which is obviously falsifiable in the first sense is never falsifiable in this second sense. (For this reason I have used the expression falsifiable as a rule only in the first, technical sense. In the second sense, I have as a rule spoken not of falsifiability but rather of falsification and of its problems)" — Karl Popper, Realism and the Aim of Science And please, don't tell me that Popper only explained that falsifiability only means that it is possible in principle whereas a falsification means that it has actually occurred. It would not have been necessary for Popper to insist on that. In the context of this quote, Popper explains that he was misunderstood by many because of this confusion between two meanings of falsifiable. Dominic Mayers (talk) 14:35, 25 January 2020 (UTC) Anyway, I can work on this article. I know very well where I am going, but I cannot do that if other editors don't agree on the basic. It would be pointless to suggest a lead that I think is acceptable as a basis for a discussion, if you don't feel the importance of this quote from Popper. The discussion has to start with some agreement on the basic. Dominic Mayers (talk) 14:41, 25 January 2020 (UTC) Firstly, I do not see a significant point of disagreement between us as to the nature of falsification. Your improvements are worthy. Secondly, working with other editors is in the very nature of Wikipedia, so if you cannot do that you are in the wrong place; but further, it seems to me that you can work with others, because you are. The first sentence is the most difficult to write. At the least, forget about it and continue with the work you have done. A Wiki article is always a work in progress. For my part I intend to add more detail on naive falsification, contrasting it with verificationism (which redirect here, so needs at least a mention) and explaining modus tollens; and that might be enough for me. Banno (talk) 15:38, 25 January 2020 (UTC) I feel that we are in agreement about falsification or falsificationism. The disagreement is about falsifiability, which is a more abstract concept. It is more difficult to conceptually understand what is falsifiability than to conceptually understand what is a logic. Even though Popper is not as formal as we are in modern mathematical logic, he reuses the concept of mathematical logic with an extra layer on top so that we have the link with observations. If you don't see that falsifiability is at this level of complexity and abstraction, then you are missing the point and you should not oppose well sourced edit by others on the subject of falsifiability in this article, especially in the lead. The first sentence in the current lead A claim is falsifiable if some observation might show it to be false. is clear and simple. We could even say that it looks good. Yes, but it's not what Popper meant by falsifiability and it is non sense as a criterion of demarcation. Dominic Mayers (talk) 20:22, 25 January 2020 (UTC) It should be noted that there is a consensus among two (perhaps three) editors here that alternatives such as: A claim is falsifiable if in principle some observation might show it to be false. A claim is falsifiable if we can conceive some observation that shows it to be false. only makes it worst. Dominic Mayers (talk) 13:02, 26 January 2020 (UTC) The approach that I propose is as follows. We write the body of the article in a way that is extremely well sourced with multiples sources. By this I do not mean that it should be superficially sourced by taking some sentences out of context, but that it should reflect the global understanding of many authors, especially Popper. This requires that we have the appropriate understanding of the subject. The body of the article should be self sufficient and not depend on a definition or anything given in the lead. This means that the lead should have no content, especially not a definition, that is not given in the body of the article. The lead should only summarize the article. This systematic approach would most likely mean that the lead will NOT start with a definition of falsifiability, but it might introduce it by saying what it is used for. I cannot say what it will be now, because the article is not written, but it might look like Falsifiability is a logic-based concept used in critical rationalism to delimitate the statements that can be empirically assessed. This sentence does not define falsifiability, but yet it says a lot more than the current sentence. It includes the so important link with logic. It refers to the empirical basis. It relates to a key point that is made in the article: ... ways in which one might assess the theory. It is less attached to the notion of falsification. We know that the corroboration of falsifiable statements is also very important. In fact, a high degree of falsifiability is considered valuable, because it makes a corroboration more meaningful. Falsifiability is a logic-based concept used in critical rationalism to demarcate the statements that can be empirically assessed from the others. Falsifiability differs from verifiability, which was held as fundamental by philosophers such as Hans Reichenbach, because in order to verify the claim "All swans are white" one would have to observe every swan, which is not possible, whereas the single observation "Here is a black swan" is sufficient to falsify it. Falsifiability was introduced by the philosopher of science Karl Popper in his book The Logic of Scientific Discovery, as an answer to both the Problem of Induction and the Demarcation Problem. He saw falsifiability as the cornerstone of critical rationalism, his theory of science. Falsifiability, as a key notion in the separation of science from non-science, has featured prominently in many scientific controversies, even being used as legal precedent. Dominic Mayers (talk) 00:32, 27 January 2020 (UTC) Issues Unlike the existing lead, it doesn't say what falsification is, in the first line. calling it "logic'based" says nothing. It has far broader application that just critical rationalism. "delimate" is obscure "empirically assessed by observations" - redundancy Falsifiability is not corroboration Why Reichenbach? "whereas the single observation "Here is a black swan" is sufficient to falsify it" - as discussed, only in naive falsification. repetition - demarcation and "separation of science from non-science" Apart from that, fine. Banno (talk) 23:29, 26 January 2020 (UTC) Response to issues: Unlike the existing lead, it doesn't say what falsification is, in the first line. The new first sentence says a lot more about falsifiability than the current first sentence. In fact, the current first sentence has a valid content only if you interpret it in the light of the new sentence. In other words, any difficulty that one has with the new sentence only shows how confusing the current sentence might be for a reader that tries to make sense of it as a criterion of demarcation. I really disagree with that. "A claim is falsifiable if some observation might show it to be false" is exactly right;it is succinct; it is easily understood. A reader trying to understand demarction might do well to continue to reading the section on demarcation. calling it "logic'based" says nothing. For you maybe, but it obviously says a lot. You cannot expect to get the whole notion in one sentence. The main idea requires the use of a logical concept that goes way beyond modus tollens. It has far broader application that just critical rationalism. I don't see what you mean. It's only about falsifiability and critical rationalism. "delimitate" is obscure I am using it to mean "define the set of". Delimitate the statements that ... = Define the set of statements that ... . We can use demarcate the statements that ... from others. This can be improved. I agree. Falsifiability is not corroboration No, but it makes corroboration meaningful. Falsifiability is not falsification either, which is a more important point to understand. Falsifiability makes logical falsification possible and corroboration interesting. Why Reichenbach? It's fine to name one typical proponent of a viewpoint. Reichenbach was very expressive on the subject. I intend to quote him to illustrate what was the prevailing view. It was already removed, but what was there was not redundant, because the observations were qualified. "whereas the single observation "Here is a black swan" is sufficient to falsify it" - as discussed, only in naive falsification. No, at the logical level, only one falsifier is sufficient. Naive falsificationism refers to the methodology. Here we say that the statements are assessed, not rejected, so we perfectly fine. repetition - demarcation and "separation of science from non-science". There is already a repetition in the current lead. Besides, some repetitions are OK. Dominic Mayers (talk) 23:45, 26 January 2020 (UTC) It's a valid point, but only superficially. The demarcation problem existed before Popper and it was not at all the same demarcation as the one defined by falsifiability. So, when we refer to the demarcation problem, we refer to something else. Still, I agree that the possible confusion between these two demarcations creates an issue. Dominic Mayers (talk) 01:02, 27 January 2020 (UTC) When we say that it's not possible to observe all swans, we are not relying on the difficulty in the observation of every swan individually. It would be still impossible, even if scientists would agree by decision on each observation. I am not sure that it will be useful to make the sentence more complicated : in order to verify the claim "All swans are white" one would have to observe every swan, which is not possible even if scientists would accept as true each individual observation, whereas the single observation "Here is a black swan", if accepted as true, is sufficient to falsify it. The current sentence is fine. Dogmatic falsificationism is about not realizing that scientists must accept falsifiers by decisions, as covered in the section "Falsificationism as a solution..." of the article. Such decisions are implicit in the sentence. So it's not dogmatic. It's not naive either, because we are not saying that the falsified statements will be rejected. It's clear in the context, that the statements are only assessed. The point is that it's impossible to consider all these subtle points in the lead. Now, in constrast, the current first sentence makes a clear and simple point, which is simply wrong, unless one guesses the interpretation that is introduced in the new first sentence. Moreover, I never mentioned it, the next sentence claims to be an example of the definition, but it's not. Dominic Mayers (talk) 01:39, 27 January 2020 (UTC) If you need to reply to a specific point that I wrote, simply quote the point like that blah blah blah (this is the point) at the bottom after my signature and then reply. Dominic Mayers (talk) 02:27, 27 January 2020 (UTC) I disagree that naming one philosopher among many that shared a similar view creates a bias. There is no bias, because it's just a way to illustrate the view. Note that Reichenbach was the leader of the Berlin circle. So, his view is representative. However, this is not an important issue in comparison with your belief that the first sentence is perfectly fine. You do realize that I cannot agree with this. Dominic Mayers (talk) 02:34, 27 January 2020 (UTC) Sure - like A. J. Ayer... or maybe Rudolf Carnap. The point being, if you name one you biase the intro; if you name enogh to make in unbiased, you lose clarity. So name noone except Popper. Being from the Berlin Circle counts agains hm, since Popper was writng in Viena - hence the Vienna Circle would be far more appropriate. What about changing it to Vienna Circle?Banno (talk) 05:10, 27 January 2020 (UTC) Yes, we can do that. Ok with Vienna Circle if you feel it's better. This issue is negligible, because it does not reflect a disagreement about the fundamental concepts that should be presented in the article. Editors does not have to agree on every thing. They do not even have to agree on all fundamental concepts. But, the article will never be nice if the editors do not agree on the fundamental concepts that should be presented and about the way to present them. So, I hesitate to discuss this minor point about how to illustrate the fact that at the time the classical view of science was verificationism. In contrast, the role of mathematical logic in Popper's falsifiability is a fundamental concept and it's important that we agree that no sentence can pretend to define falsifiability or say the essential about it and yet say nothing about the role of mathematical logic. For me, it should be presented before or together with any fundamental sentence about falsifiability in the article. Dominic Mayers (talk) 16:05, 27 January 2020 (UTC) You wrote: I really disagree with that. 'A claim is falsifiable if some observation might show it to be false' is exactly right;it is succinct; it is easily understood. A reader trying to understand demarction might do well to continue to reading the section on demarcation. I have written a long text of 7 paragraphs to explain why it's not only a question of how much precise we want to be. I will try to summarize, but I am afraid it will take a few paragraphs anyway. The main thing we want to say about falsifiability is that it implies that we can empirically assess the theory. So, we want a simple criterion that guarantees that. The criterion "some observation might show it to be false" does not say that because it rejects valid theories, unless of course by "observation" we mean an abstract statement that scientists agree is empirical, what Popper calls a basic statement. But, if this is what we mean by "observation", then we should say it and the issue will disappear. If observations are not abstract basic statements, the natural interpretation of this simple sentence is simply that somehow the law is contradicted and so it is false. This interpretation clearly does not work, because it rejects valid theories. Please, see that there is no way out of this. If you have any other interpretation of this simple sentence, which works and is not in terms of basic statements, write a paper about it and you will become a celebrity. So the first sentence in the lead is a misleading definition. This paragraph is about the importance of falsifiability as a distinct concept. If we don't appreciate falsifiability as a separate concept in itself, different from falsificationism, then the previous paragraph as little consequence and appears as a moot point. Lakatos has shown that a lot can be understood and explained about falsificationism without using this criterion. This means that it is possible to consider the subject of falsifiability as a detail, a technicality within the subject of falsificationism, at the least the part that can be described without it. Fine, but we need to respect that many people, including Popper, felt that this criterion was very useful. So, if we decide to focus on falsifiability, not falsificationism, but on the technical detail, which is falsifiability, then we have to be consistent with our choice and truly give the criterion. In the article, we mention falsifiability, not falsificationism, in the first sentence of the lead. We do that as if we think this criterion, this technical detail, is important. In the light of the previous paragraph, we should explain what it is in a way that works. This paragraph explains the criterion that works. If we don't understand the criterion that works, then we might continue to appreciate the criterion that does not work, because we cannot compare it to anything. To make it work, we first describe theories separated from observations, in a purely logical manner. This means that we have a language to write basic statements, universal laws, etc. and we can prove things using deductive logic. To make this abstract language empirical, but without fixing any specific theory yet, we exhibit what objects in the language correspond to physical observations, to observational data. This is a decision made by scientists. It's the first decision in Lakatos language (even though Lakatos refer to these decisions without defining falsifiability). It's the only decision that is needed to determine what are the basic statements and falsifiability, as defined by Popper. If one proposes a theory in this language, he must find basic statements that contradict the theory to show it falsifiable. The key point is that this underlying language is empirical, but yet allows all kind of "observations". In fact, if the theory is interesting, given some initial conditions, only a few basic statements that respect these conditions are true. The others are false. For example, if the law is all swans are white, under the condition that some object is a swan, the statements that this object is yellow, blue, green , black, etc. are all false. This works, but only because we use tricks from mathematical logic. It's important. Now, one might argue that it's impossible to formalize science in terms of mathematical logic, because when we write the laws of science we rely a lot on our natural language. This is true. However, Popper replies that what we learn from mathematical logic about truth, the interpretation of a language, etc. apply also to our natural language [in the context of science]. So, he reuses the concept of mathematical logic to define falsifiability. In mathematical logic, there is no proof that a logic can be used to represent our day to day mathematics. Similarly, Popper does not prove that his notion of basic statements, etc. are a correct representation of our day to day scientific theories, but it makes sense. Besides, there is no way to do some serious philosophy of science without accepting that. This is so fundamental. Dominic Mayers (talk) 08:47, 27 January 2020 (UTC) I am only trying to be practical and write something useful, not misleading, which introduces the subject as it is. What is the subject in the case of falsifiability? We just said it: it's a criterion that says that we can empirically assess the theory. My feeling is that only stating that would be a bit empty. We are going to give the correct criterion in the article, so we should say something in that direction. That's why I consider that the lead should mention that is based on logic. So, the first sentence in the lead could say that Falsifiability is a criterion inspired by mathematical logic to demarcate empirical statements from non empirical statements. This sentence says a lot. It says that it is inspired by mathematical logic, which states the correct level where the concept can be understood. It says that it is about the empiricity of the statements. The question is how to fit in the important "possibly false" requirement of falsifiability, when clearly we also want to accept "definitivel true" statements. I think we have to say it incorrectly - better to say it incorrectly than not saying it at all, but we say it as such. The next sentence would be: Informally, at the cost of some contradiction, a statement is falsifiable if some observation might show it to be false. The next thing that needs to be addressed in the current lead is that the example that is given For example, in order to verify the claim "All swans are white" one would have to observe every swan; but the observation of a single black swan would be sufficient to falsify the claim. is very interesting, but does not really illustrate falsifiability. It illustrates why it is better than verifiability, which is fine. It goes immediatly to the crux of the matter, but it must be introduced accordingly. Even before that, let us move on top the example that does illustrate falsifiability. So, the next sentence (taken from the current lead) would be: For example, "All swans are white" is falsifiable because the observation "Here is a black swan" shows it to be false. Now, we introduce and present the next example: Falsifiability differs from verifiability, which was held as fundamental by many philosophers such as those of the Vienna circle, because in order to verify the claim "All swans are white" one would have to observe every swan, which is not possible, whereas the single observation "Here is a black swan" is sufficient to falsify it. The remainder of the current lead can stay the same, except for a small change, since we already said that it is a demarcation criterion: "is a key notion .... and as such ... " becomes ", as a key notion ..., ..." As mentioned in the above discussion, the word demarcation is used again, but it's not the same demarcation. The demarcation problem refers to a separation between anything that can be useful in science and what is totally useless or meaningless. Falsifiability creates a different demarcation. So, putting all this together, the final lead would be: Falsifiability is a criterion inspired by mathematical logic to demarcate empirical statements from non empirical statements. Informally, at the cost of some contradiction, a statement is falsifiable if some observation might show it to be false. For example, "All swans are white" is falsifiable because the observation "Here is a black swan" shows it to be false. Falsifiability differs from verifiability, which was held as fundamental by many philosophers such as those of the Vienna circle, because in order to verify the claim "All swans are white" one would have to observe every swan, which is not possible, whereas the single observation "Here is a black swan" is sufficient to falsify it. Falsifiability was introduced by the philosopher of science Karl Popper in his book The Logic of Scientific Discovery, as an answer to both the Problem of Induction and the Demarcation Problem. He saw falsifiability as the cornerstone of critical rationalism, his theory of science. Falsifiability, as a key notion in the separation of science from non-science, has featured prominently in many scientific controversies, even being used as legal precedent. Dominic Mayers (talk) 19:00, 27 January 2020 (UTC) Thanks for this. I could accept "Informally", with some reservation. I've objections to "Falsifiability is a criterion inspired by mathematical logic to demarcate empirical statements from non empirical statements" being the first line. The implication is that demarcation is the be-all of falsification; it isn't. Philosophicaly, the solution to induction is at least as interesting. SO what about, as a compromise, Put simply, a statement is falsifiable if some observation might show it to be false. For example, "All swans are white" is falsifiable because the observation "Here is a black swan" shows it to be false. Falsifiability is a criterion inspired by mathematical logic to demarcate empirical statements from non empirical statements. ...leaving your footnote in as a compromise. If you are ageeable, please make the changes. Banno (talk) 21:21, 27 January 2020 (UTC) If the two statements were at the same level of specifity, changing the order would be perfectly fine, in fact, more logical. But this is not at all the situation. The statement "Falsifiability ... inspired by mathematical logique ..." sets the general context. In contrast, the statement "... some observation might show it to be false" is not the definition and is only useful to convey a specific point: the requirement that somehow the statement is falsified by observation. It's an important point. I am sure this why you focus a lot on it and feel that it's the whole thing, but it's not. So, it's more logical to start with the general point and then give the sentence that conveys a central, but yet specific point. You agree with the principle, because you mentioned that you don't want too specific at the start. I address your objection below. Before I address your objection, did you read #The_disagreement_about_the_first_sentence_is_a_fundamental_one? If we had an agreement about the fundamental aspects, we would not have this discussion. The issues that we have regarding the order of sentences, mentioning the fact that there is a contradiction, etc. are not only issues about how to present a shared understanding. They reflect a disagreement about the fundamental aspects of falsifiability. You obviously did not accept that falsifiability is fundamentally a concept inspired by mathematical logic. You are still thinking that the simple sentence "... might show it to be false." almost explain the whole thing. It will be very difficult to write a nice article in this situation. There is no problem when editors are proponents of different views with different important and notable sources, because then we present the different views each with their sources. Here, we have a disagreement on the fundamentals of a unique subject: falsifiability. Anyway, I can nevertheless reply to your objection regarding the sentence "Falsifiability ... inspired by mathematical logique ...". You wrote: I've objections to 'Falsifiability ... inspired by mathematical logic...' being the first line. The implication is that demarcation is the be-all of falsification; it isn't. Philosophicaly, the solution to induction is at least as interesting. You confuse two demarcations here. The demarcation that was a problem before Popper and the demarcation defined by the demarcation criterion, which is never presented as a problem. Clearly, a statement of the form "The [demarcation] criterion is used to demarcate ..." is totally general. Honestly, I was expecting that the critic would be that it is too general. We can find another word than "demarcate", if you feel that it might sound better. The point is that falsifiability is a criterion, i.e., a property on statements. Any property on a set, it does not matter what kind of set, set of statements or whatever, separates the set in two: the elements that respects the property and those that don't. It's well known that the statements that respect the criterion are said to be empirical. It's such a general point. Really, this statement fits perfectly in the first line. Dominic Mayers (talk) 22:55, 27 January 2020 (UTC) Think I said before that you seem to me to be overthining the issues. Critical Rationalism is only a part of the story, and we must avoid treating it as a whole. Demarcation stands beside, not ahead, of the problem of induction. The statement "... some observation might show it to be false" is the definition of falsifiability, if one accepts that there will be a bit of finessing in the body of the article. "Demarcate is to be prefered. We are writing an encyclopedic article, not an accademic paper. Would that there were more voices here. Banno (talk) 00:03, 28 January 2020 (UTC) I cannot make any sense of what you wrote here. You did not understand at all that the demarcate in the first sentence is simply a consequence that falsifiability is a property. Any property whatsoever creates a demarcation. We can reformulate the sentence so that demarcate is not used. The sentence is behind both the demarcation problem and the problem of induction, because it is as general as it can be about falsifiability. If we are going to call for other voices, let's make it clear and also call for experts. Dominic Mayers (talk) 00:13, 28 January 2020 (UTC) That'd be becasue technically demarcation is an imperative plonked on top of a bit of logic, and not "simply a consequence that falsifiability is a property". But arguing the point is not of interest to me; not here. I'm not too unhappy with the article as it stands, but would resist any attempt to make it entierly about critical rationalism. I've asked for comment on the philosophy project; but I suspect that's a bit stagnant at the moment. Are there folk you might bring in? Banno (talk) 00:36, 28 January 2020 (UTC) You really don't get it. You keep responding as if the word "demarcation" which was there in the first sentence was the demarcation in philosophy. If I write "Primality is a criterion to demarcate the prime numbers from the others", do you conclude that I refer to the demarcation problem in philosophy? Of course, you don't. Exactly in the same way, in the sentence "Falsifiability is a criterion to demarcate empirical statements from the others.", we do not refer to the demarcation problem in philosophy. There is a link, in this case, but there is also a link with the problem of induction. It's as much linked with one than with the other. You are too attached to the choice of words. Dominic Mayers (talk) 01:09, 28 January 2020 (UTC) Informally, a statement is falsifiable if some observation might show it to be false. For example, "All swans are white" is falsifiable because the observation "Here is a black swan" shows it to be false. The apparent contradiction seen in the case of a true but falsifiable statement disappears in the formal definition, which is inspired by mathematical logic. Falsifiability differs from verifiability, which was held as fundamental by many philosophers such as those of the Vienna circle, because in order to verify the claim "All swans are white" one would have to observe every swan, which is not possible, whereas the single observation "Here is a black swan" is sufficient to falsify it. It was introduced by the philosopher of science Karl Popper in his book The Logic of Scientific Discovery, as an answer to both the Problem of Induction and the Demarcation Problem. He saw falsifiability as the cornerstone of critical rationalism, his theory of science. As a key notion in the separation of science from non-science, it has featured prominently in many scientific controversies, even being used as legal precedent. Dominic Mayers (talk) 00:30, 28 January 2020 (UTC) empiricalness? Banno (talk) 00:38, 28 January 2020 (UTC) I took it out (before you wrote this), but it is a word sufficiently used to appear in wiktionary. So it was perfectly fine. A language evolves and new words are used. The meaning was not hard to guess. I took it out, because I decided to look for a way to move the mention of the correct definition later. It's very important to mention very early the issue with the naive sentence, because, maybe not for you, but for many people, 10%,5%, 20%, I have no idea. This sentence is confusing. Dominic Mayers (talk) 00:53, 28 January 2020 (UTC)

Technical article or brief jump-off point?

In this January 25 edit in this talk page, we have: Excellent. I guess we proceed by working out what belongs here, and what belongs in critical rationalism. I think this is the place for a discussion of the logic of falsifiability, and a place to contrast it with verificationism. Overall, I agree that this article ought be a brief jump-off point to a wider understanding of the topic. I do not see it that way. Yes, we should try to separate what belong in falsifiability from what belong in critical rationalism, but my understanding is that falsifiability, because it is a technical definition, is the opposite of a jump-off point.

Dominic Mayers (talk) 19:49, 28 January 2020 (UTC)

Demarcation

Overall the changes of the last few weeks have been positive. The Demarcation section is an exception. The first two paragraphs ramble on about verificationism as a theory of meaning, which is all but irrelevant here;Third paragraph is ambiguous; and no where does it tell me what Demarcation actualy is. Banno (talk) 21:18, 30 January 2020 (UTC)

No. On the contrary, the fact that a theory of meaning was a way to separate science from non science is the whole point. This is the way Popper and many others explain it. Dominic Mayers (talk) 23:22, 30 January 2020 (UTC)

I intend to complete the section about induction with Popper's solution as he presents it himself. Currently, this section presents Lakatos view only. Falsifiability could be defined here. So, the demarcation section will naturally refers to it. I think you keep doing the same mistake: you confuse the demarcation created by falsifiability with the demarcation in the demarcation problem. Falsifiability demarcates between empirically meaningful statements and the others, but the demarcation that was attempted before Popper was completely different and this created a problem. This is what the demarcation problem section should be about. It's ok that falsifiability is defined before, because falsifiability is also important in the problem of induction. Dominic Mayers (talk) 23:09, 30 January 2020 (UTC)

Suggests here what you have in mind. You might have your own view. Please share it before starting to edit, especially if it means a complete reorganization. Because, I might have to undo a lot of it and it's not time well spent for you and me. Dominic Mayers (talk) 22:30, 30 January 2020 (UTC)

We must consider the growth of knowledge

We must take into account that the problem of induction and the demarcation problem are facets of the main problem of epistemology: the growth of knowledge. This is Popper's view on epistemology and it's the relevant view.

The positivists saw the inductive methodology as an explanation of the growth of knowledge and at the same time they demarcated what was considered scientific. The inductive methodology was an attempt to explain growth within the demarcated scientific theories. They provided a wrong solution to the correct problem. Falsifiability is Popper's solution to the same problem.

We cannot give a solution to the demarcation problem without explaining how the demarcated scientific theories grow. We cannot give a solution to the problem of induction without explaining how it is consistent with a growth of knowledge. It is also useful to explain how falsifiability compares to other attempts to a solution. This is the way the article considers the demarcation problem, for example.

Most if not all critics turn around the problem of the growth of knowledge. So, it will be totally inappropriate to present these critics without presenting falsifiability accordingly.

Dominic Mayers (talk) 01:44, 31 January 2020 (UTC)

citations for naive falsification

Naive Falsification is a term used by Lakatos in "Criticism and the methodology of scientific research programs": Proceedings of the Aristotelian Society, New Series, Vol. 69 (1968 - 1969), pp. 149-186 Published by: Blackwell Publishing on behalf of The Aristotelian Society

"For (Popper) the idea that on 'refutation' one can demand the rejection, the elimination of a theory, is 'naive' falsificationism" (p.150)

and then, explaining why the term is significant,

"In this paper I shall first show that in Popper's philosophy two different positions are conflated. Kuhn understands only Popper, the naive falsificationist (I shall call him Popper,),and his criticism of Popper1 is correct. I shall even strengthen it. But Kuhn does not understand a more sophisticated Popper-Popper2- whose rationality goes beyond naive falsificationism. I shall try to explain Popper2's position and strengthen it, mainly by stripping it of naive falsificationism. This improved Popperian position may escape Kuhn's strictures and provide a rational explanation of scientific revolution." (p. 151)

This paragraph has a pivotal place in the story of falsification, and ought either be included or summarised. Banno (talk) 06:26, 25 January 2020 (UTC)

The article already mentions the fact Lakatos presented naive falsificationism as a way to better explain Popper's falsificationism. However, the paragraph that you cite introduces a thesis that actually critics Popper. See in particular pages 159-167. The article should say something about this. It was at the source of a very serious dispute between Popper and Lakatos. I have excellent sources about this. However, it makes absolutely no sense to go into this without giving the definition of falsifiability, because Popper's angle cannot be understood without it. We covered a lot without using the definition, but here is a central point that cannot be covered without it. We could give Lakatos's angle only, but that would not be adequate and only contribute to the confusion. Dominic Mayers (talk) 17:43, 29 January 2020 (UTC)

@Dominic Mayers: Congratulations on having made the article much much clearer and better overall! --Macrakis (talk) 19:22, 29 January 2020 (UTC)

Thank you! Dominic Mayers (talk) 03:18, 31 January 2020 (UTC)

Errata: The article you cite is very similar to the first chapter of "The methodology of scientific research programmes" (1978), which contains the same first sentence that you cited. When I mentioned pages 159-167, I meant in this book, not on the paper that you cited. Dominic Mayers (talk) 03:17, 31 January 2020 (UTC)

Formalizing Modus tollens and link to fallacies are not useful

To understand falsification, it's not useful to formalize modud tollens as the formal rule

${\frac {P\to Q,\neg Q}{\therefore \neg P}}$

Doing that, will only make the reader, which is lost in the technicalities, believe that he needs to understand this formalization. It might just scare him away even more. In fact, even mentioning the rule informally might not be needed, because it's very intuitive. The reader can get confused if we start to explain modus tollens as if this was what a key concept in falsifiability, when in reality what the reader needs to know about it is nothing else than the intuitive notion, which is a basic requirement for the understanding of anything in science, mathematics and epistemology. I am not saying that knowing how to formalize a logic, prove equivalence between logical systems, etc. is a basic requirement. I am saying that this is not what falsifiability is about and let's not confuse the reader about this.

Dominic Mayers (talk) 05:45, 28 January 2020 (UTC)

I added this paragraph:

Falsifiability uses the valid inference modus tollens: if from a statement $P$ (say some law with some initial condition) we logically deduce $Q$, but what is observed is $\neg Q$, $P$ is false. For example, given the law "all swans are white" and the initial condition "there is a swan here", we can deduce "the swan here is white", but if what is observed is "the swan here is not white" (say black), then "all swans are white" is false, or it was not a swan.

But I am not convinced it's useful. As I explained above, it might confuse the reader by letting him believe that falsifiability is about this kind of things, which belong to basic logic. There are two cases. First case: the reader is not good in basic logic (but still he naturally uses it). In this case, the reader may start to think that the issues in falsifiability are at this level, which is wrong. Second case: the reader knows about basic logic. In this case, the reader does not need this paragraph and may even find it annoying. Well, there is a third case: the reader cannot apply even intuitively basic logic. In this case, the reader needs more than a single paragraph.

Dominic Mayers (talk) 17:11, 29 January 2020 (UTC)

For the same reason, a link to a fallacy such as affirming the consequent is not useful, especially when it does not apply. I refer to this sentence in the article :

We cannot validly argue from "here is a white swan" to "all swans are white"; doing so would be to affirm the consequent.

This is not what falsifiability is about. The positivists with their inductive methodology understood very well the rules of logic. It is misleading for the reader to put an emphasis on fallacies. Moreover, in this particular case, the link points to a fallacy that does not even apply at all. There is no implication from "All swans are white" to "Here is a white swan", because "All swans are white" is equivalent to "There is no white swans or all of them are white". Therefore there is no antecedent to conclude from an affirmation of the consequent - there is no antecedent and no consequent in fact.

There might be a named fallacy that applies here, but it's not important, because it would be inappropriate to suggest that the great thinkers of the Vienna Circle were using fallacious logical reasoning. That's not what this article is about. We need to consider the fundamental philosophical issue that these philosophers had to face. Inductive reasoning was even considered as valid by Hume. Because he knew very well that it was not possible within the range of logic, he simply decided that it was valid in the psychological area. Even recently great thinkers have considered that we might need to reintroduce induction to explain the progress of science. Popper's thesis that it's not even a part of our day to day way to acquire knowledge is amazing.

Dominic Mayers (talk) 21:38, 30 January 2020 (UTC)

including links that explain the logical structure of a logical process is useful. Leave it in. Banno (talk) 22:10, 30 January 2020 (UTC)

At the least, find the correct fallacy that makes one go from one (or more) instance(s) to an universal law. I don't think it exists. Or if it exists, it's called the induction fallacy. I will agree with a few links, but we don't want to say too much, because it will be misleading. Dominic Mayers (talk) 22:24, 30 January 2020 (UTC)

CItation provided. Take it up with Grayling. Banno (talk) 03:11, 31 January 2020 (UTC)

Alluding to the provided citation, you wrote "as if it were needed." It was indeed not needed for me to lookup your citation to know that it could not verify the added sentence "doing so would be to affirm the consequent". But, I did lookup, just to go with the rule and, as expected, it does not. We see "affirming the consequent" (at page 457, not 397), but the implication in the text is not from a universal law to an instance. So, you have not at all verified the content.

We have a problem here. The fallacy in the sentence, if we want to see it as a fallacy, is clearly an inductive fallacy. I provided a reference to go along with the principle of mentioning a fallacy. If this does not work, we have a serious problem. I made a very sincere compromise in accepting to point out a fallacy. Even if Popper mentioned a fallacy somewhere, it's obvious that it's not a central point. Dominic Mayers (talk) 03:46, 31 January 2020 (UTC)

Away from naive falsificationism

Here is a rough draft for this subsection:

Naive falsificationism is the claim that methodological falsifications can by themselves explain how we gain scientific knowledge. Very often we must deal with two or more competing theories which can both be non-falsified (say by not making the proper decisions). Considering only falsifications, we can not explain why we pick one theory above the other. We can not explain why often a corroborating experiment is seen as a sign of progress. Popper's falsificationism or simply falsificationism use both falsifications and corroborations to explain progress in science. How corroborations and falsifications can explain progress in science was a subject of disagreement between many philosophers, especially between Lakatos and Popper.

Popper clearly distinguished between the creative and informal process from which theories emerge and the logical and formal process where theories are falsified or corroborated. The main issue is whether the decision to select a theory among competing theories in the light of falsifications and corroborations should be moved in the logical part as some kind of formal logic. It is a delicate question, because this logic could be considered as inductive: it selects a universal law in view of instances. The answer of Lakatos and many others to that question is that it should. In contradistinction, for Popper, the creative and informal part is guided by methodological rules, which naturally say to favor theories that are corroborated, but this methodology can hardly be made rigorous. Popper does not mention corroborations when he describes the purely logical part.

Popper's way to analyze progress in science was through the concept of verisimilitude, a way to define how close a theory is to the truth, which he did not consider very significant, except (as an attempt) to clarify a concept already clear in practice. Later, it was shown that the specific definition proposed by Popper cannot distinguish between two theories that are false, which is the case for all theories in the history of science. Today, there is still on going research on the general concept of verisimilitude.

Dominic Mayers (talk) 03:27, 2 February 2020 (UTC)

I will add references later. All of this can be very well sourced. Dominic Mayers (talk) 03:27, 2 February 2020 (UTC)

The technical tag at the top of this talk page

There is a difference between simplistic and simple. We need to make sure that the technical tag is not used because we have a too simplistic view. From a simplistic view, even an article that is as simple as it can be would appear too technical. From the discussions that we have above, I cannot rule out the possibility that there is an expectation for something too simplistic. Anyone that feels the tag is still needed, please explains what is your view of the subject: how the article could be made less technical, without becoming simplistic. Dominic Mayers (talk) 01:27, 8 February 2020 (UTC)

The third sentence The apparent contradiction seen in the case of a true but falsifiable statement disappears once we know the technical definition. might illustrate very well the situation. There is nothing technical in this sentence. The apparent contradiction in "true, but falsifiable" is obvious. Mentioning that a technical definition can take care of this is also not technical at all. The sentence is not technical. A reference to the existence of a technical definition might be incompatible with a "simple" view of the subject, but that is a different issue and, in that case, we need to rule out the possibility that this "simple" view is too simplistic. Dominic Mayers (talk) 01:44, 8 February 2020 (UTC)

Section Criticisms

The section Criticisms needs to be entirely revised. It's not well sourced and most of it is misleading. Kuhn himself explains in great details how much his view of Science is close to Popper's view. In particular, he says that his view about revolution in Science is the same as Popper - this can be sourced. He basically has two criticisms. The first is that Popper does not describe normal science, only the big revolutionary steps. Popper agreed with him and acknowledges that he has not thought about that (but adds that normal science is not interesting...) The second criticism is already mentioned in the article and it's the same as Lakatos's criticism.

What is written about Feyerabend seems correct, but it seems more a criticism of Lakatos's sophisticated falsificationism than a criticism of Popper's falsificationism. This has to be clarified using sources.

I am not sure about Socal and Bromont. It's not at the same level. Their criticism is directed on naive falsificationism or at best they repeat the old criticism of Lakatos and Kuhn. Here is how The Guardian quoted them :

"When a theory successfully withstands an attempt at falsification, a scientist will, quite naturally, consider the theory to be partially confirmed and will accord it a greater likelihood or a higher subjective probability… But Popper will have none of this: throughout his life he was a stubborn opponent of any idea of 'confirmation' of a theory, or even of its 'probability' … [yet] the history of science teaches us that scientific theories come to be accepted above all because of their successes."

— Sokal and Bromont, Fashionable Nonsense

Because they are themselves fashionable, it may be noteworthy, but it's not at all at the same level as Kuhn and Feyerabend.

Dominic Mayers (talk) 03:00, 5 February 2020 (UTC)

Testing on mobile... There is a bug. I was testing a response to the last contribution, not here.Dominic Mayers (talk) 15:53, 9 March 2020 (UTC)

Second test. I see clearly that I am in the section "The lead is inadequate" and I try to respond. I am on my mobile phone. Dominic Mayers (talk) 16:02, 9 March 2020 (UTC)

The lead is inadequate

This lead was written before important additions to the article. These additions make clear that the basic concept of falsifiability depends on a separation between the formal/logical and creative/subjective aspects of science. Falsifiability exists in the formal/logical side only. The creative/subjective aspect is only used to guess new theories and define the empirical basis, which is decided by conventions, not by logical rules. This separation is central in almost all sections, including the section about controversies. It's weird that the lead says nothing about it and its first sentence even contradicts this central idea. The informal definition mixes both aspects and thus totally fails to convey the main point. We have to work on this. It's ok to be informal in a brief summary if it conveys the central idea, but that's not what we have in the current lead. Dominic Mayers (talk) 14:59, 5 March 2020 (UTC)

Hi. This sounds very important but I'm not sure I understand the distinction. Would you help me? BillyGoatsGruff2020 (talk) 07:48, 9 March 2020 (UTC)

The distinction is a very well verified part of the article. Have you tried to read the sources? If you need help to get the sources, I can send you electronic version (when it's possible), at the least some sections. I am not fanatic at all about always referring to the sources all the times in the discussion page. On the contrary, I believe it's necessary that editors exchange their respective understanding in a very natural manner, without having to mention the sources all the times. However, you need to have some understanding of the sources.

The distinction is central in Popper's thinking. You can see it in The Logic of Scientific Discovery, Realism and the Aim of Science, etc. Thornton wrote : "Popper has always drawn a clear distinction between the logic of falsifiability and its applied methodology. The logic of his theory is utterly simple: if a single ferrous metal is unaffected by a magnetic field it cannot be the case that all ferrous metals are affected by magnetic fields. Logically speaking, a scientific law is conclusively falsifiable although it is not conclusively verifiable. Methodologically, however, the situation is much more complex: no observation is free from the possibility of error—consequently we may question whether our experimental result was what it appeared to be" and I am not saying that Thornton covered the whole subject with this sentence about experimental errors. The informal aspect of science shows up even more in the way we propose theories. This is perhaps where the distinction is most important. Dominic Mayers (talk) 19:47, 9 March 2020 (UTC)

Attempt to improve the lead

I reverted this attempt to improve the lead. The first sentence in this attempt was:

Falsifiability is the assertion that any hypothesis must be inherently disprovable before it can become accepted as a scientific hypothesis or theory.

Falsifiability is simply not an assertion. Right there, we already have a problem. But, let's assume the goal was to use "falsifiability" to refer, in a very broad manner, to what is asserted by its associated philosophy. The philosophy asserts that a statement must be falsifiable before it can become accepted as a scientific theory. So, this attempt would make sense, if "inherently disprovable" was the definition of falsifiable. Otherwise, it's not even the correct philosophy. I never seen "inherently disprovable" as the definition of falsifiability.

We should not have to discuss the content of the lead in terms of its direct verifiability in the literature. What is important is that it summarizes the key points of the article. We don't say in the article that falsifiable means inherently disprovable. I understand the intuition though: if it is falsifiable, there must be some way to prove it false. Yes, but unless we explain correctly the way of proving the statement false, we miss the key point. The article explains it. The way is through the notion of empirical basis. The empirical basis is one of the most important key point in the notion of falsifiability.

Dominic Mayers (talk) 18:09, 14 March 2020 (UTC)

But this attempt was actually useful. It made me realize that a simple key point to add in the lead is that the difference between the informal and technical definitions is only the definition of the empirical basis, i.e., the basic statements that can be used to show a statement false. Based on this I made this edit. Dominic Mayers (talk) 18:50, 14 March 2020 (UTC)

Non verified and, I believe, non verifiable statement

It just occurred to me that the following statement in the article is not verifiable.

These observation or basic statements constitute the empirical "basis".

For Popper, a basic statement does not have to be an actual observation. This contradicts what is written just before in the article:

the written form of the observation matches with an actual observation. 

Popper so much insisted that a falsifier does not have to be actually possible that I consider this a very serious unverifiability issue. It's not entirely surprising, because I and others were aware that we have not given the actual definition of falsifiability in the article. We just pointed to it quickly in the lead. Personally, even though I know that my personal view does not really matter, I would like to say that I find interesting that we could go a long way, following Lakatos, without actually defining falsifiability. Well, it's not entirely true. The definition is implicit in the decisions (mentioned by Lakatos) that must be taken by the scientists. We only need to relax the decision that is mentioned above. The observation needs only to refer to existing technologies. It can combine different parts that are possible separately, but not possible together. I remember that Popper discusses that somewhere. For example, identifying a raven refers to existing technologies and observing the color red also, but observing a red raven is impossible (at the least, it is what we believe) and yet it is a valid falsifier. In contrast, "all men are mortal" is not falsifiable because no technology can allow us to conclude that a man is immortal. Dominic Mayers (talk) 01:36, 7 April 2020 (UTC)

The picture on the left illustrates another example, but in this case it is the initial condition in an eventual falsifier, observing an angel, that cannot be supported by conventionally accepted technologies. We have already given an example (the red raven) that illustrates that a falsifier does not have to be possible. In a footnote, we have another example: "this duck weigh two tons" is impossible and yet a valid falsifier for "all ducks weigh less than one ton." The point is that we have many examples that illustrates that the empirical basis used to falsify the theory can contain impossible falsifiers, but these falsifiers, including the impossible ones, must be describable in terms of conventionally accepted technologies. I mention this because what is an empirical basis is the central notion in the definition of falsifiability. Once we have the empirical basis, falsifiability can easily be defined in a rigorous manner. Dominic Mayers (talk) 10:24, 7 April 2020 (UTC)

Planned modifications to correct for a missing definition

List of issues associated with a missing definition (the proposed solutions will follow):

• Issue 1. We do not actually give the full solution in the section "Critical Rationalism as a solution to the problem of induction".
• Issue 2. The sentence that I emphasized in this excerpt

  ... statements of observations, such as 'this is a white swan'. Logicians describe such statements in terms of existential quantification, since they assert the existence of some particular thing.

  is incorrect in the following sense that statements such "this swan is white" can as well be expressed in terms of universal quantification. Popper gives the following example:

  For example, the statement ‘Napoleon was born in Corsica’ can be expressed in the form, (x) (x = N → φx), in words: it is true for all values of x that, if x is identical with Napoleon, then x was born in Corsica.

  — Karl Popper, Logic of Scientific Discovery, sect. 15

  To complement this, let me add that Popper attributes no importance to whether we identify a singular entity with an expression such "this swan" or by use a proper name such as "Napoleon". In both cases, he says, we refer to a singular object. Popper wrote an entire section and more to emphasize that purely logical statements, which he calls strict statements, are not able to express in themselves the notion of singular objects. We need the separate notion of an individual constant, such as expressed by the symbol N for Napoleon. The same issue shows up in this sentence

  From an existential statement to a universal statement?

  Given the previous context, this sentence confuses (strict) existential statements with statements of (singular) observation. No one, not even the positivists of the Vienna Circle, have considered the possibility of passing from an existential statement to a universal statement. What was considered is how to pass from a collection of observation statements to a universal statement.
• Issue 3. For completeness, I mention that the following sentence is incorrect and the reference given does not verify it.

  We cannot validly argue from "here is a white swan" to "all swans are white"; doing so would be to affirm the consequent.[3]

  The fallacy affirming the consequent given in the reference is: p ⟶ q & q; therefore p. In our case, we observe a fallacious move from q="here is a white swan" to p="all swans are white". Unfortunately, this cannot be directly explained by "affirming the consequent", because we don't have the required p ⟶ q. We must consider the intermediary statement p'="all swans are white and here is a swan". Only then we have the required p' ⟶ q. Indeed, the fallacious move from q to p', i.e., affirming the consequent, logically supports the fallacious move from q to p, because p' ⟶ p. But, even if a fallacious move from q to p' logically supports a fallacious move from q to p, when we observe a fallacious move from q to p, we cannot affirm that there was a fallacious move from q to p'; funnily, this would be similar to affirm the consequent.
• Issue 4. I recall without further details the issue mentioned in #Non_verified_and,_I_believe,_non_verifiable_statement. The following statement in the article is not verifiable.

  These observation or basic statements constitute the empirical "basis".

• Issue 5. The first two paragraphs in section Falsifiability#Demarcation_problem discuss more the solution, falsifiability, than the problem. They refer to falsifiability, which is not yet well defined, because the empirical basis is not.
• Issue 6. Similarly, the paragraph "Demarcation_and_growth_of_knowledge" in the same section would fit better after we have defined falsifiability.
• Issue 7. The section Falsifiability#Demarcation_problem does not properly present Popper's view on the demarcation problem, which unfolded when he realized that some theories where not falsifiable in contrast with Einstein's theory of general relativity.
• Issue 8. Of course, the main issue is that the notion of empirical basis and the definition of falsifiability are missing.

Solutions:

• Solution 1. The title will be changed to "The problem of induction."

• Solution 2. The entire paragraph is not really needed for the remainder of the section. It can be seen as a failed attempt to introduce the empirical basis, which would be easier to introduce later after the explanation that decisions must be taken by agreement between scientists in a falsification process. We will reuse parts of this paragraph when we explain the empirical basis later.

• Solution 3. Not yet clear. A statement that is verified by the reference is

  We cannot validly argue from "here is a white swan" to "all swans are white (and here is a swan)"; doing so would be to affirm the consequent.[3]

  However, even though this modified statement is now verifiable I think it is misleading to suggest that those who believe in the induction principle are making such a fallacy. Yes, there is a reference that mentions this fallacy in the same context, but it's not the common view. Never seen this elsewhere. It is problematic. It suggests that the fallacious logic is as follows: we go from "here is a white swan" to "all swans are white (and here is a swan)" and then to "all swans are white". It could have been some other wrong principle such as an induction principle, which does not require an unnatural intermediary statement.

• Solution 4. This sentence and the two subsequent sentences (i.e. the remainder of the paragraph) can be taken out. We can reuse part of these sentences later when we define the empirical basis.

• Solution 5, 6 and 7. We rewrite the section Falsifiability#Demarcation_problem using the paragraph that refers to the Vienna circle, but also add the independent perspective that Popper had when he considered Marxism, introspective psychology, etc.

• Solution 8. After the section Falsifiability#Demarcation_problem, we refer to Popper and many other references to give the definition of falsifiability. It will not be so difficult to explain the empirical basis, because the need for decisions by agreements between scientists is discussed in previous sections. The discussions that we removed from section Falsifiability#Demarcation_problem can naturally follow the definition.

Dominic Mayers (talk) 20:55, 8 April 2020 (UTC)

The empirical basis and falsifiability

I have implemented solutions 1, 2, 3, 4, 5, 6 and 7. I think that there is now a good context to do step 8, which is the definition of falsifiability and of the empirical basis. I propose to call the new section "The empirical basis and falsifiability" or "Falsifiability and the empirical basis". Dominic Mayers (talk) 16:08, 11 April 2020 (UTC)

Texts that might be reusable

First, here are texts that were taken out and that might be reusable in the section:

Popper noticed that two types of statements are of particular value to scientists. The first are statements of observations, such as 'this is a white swan'. Logicians describe such statements in terms of existential quantification, since they assert the existence of some particular thing. For Popper, such statements form the empirical basis of scientific theory. The second type of statement of interest to scientists categorises all instances of something, for example "All swans are white". Logicians describe these in terms of universal quantification.

This exerpt was taken out from the section "The problem of demarcation":

Given an empirical basis, the criterion of demarcation draws a sharp line between those theories that are scientific, and those that are un-scientific. It is useful to know if a statement or theory is falsifiable, if, for no other reason, than that it provides us with an understanding of the ways in which one might assess the theory. One might at the least be saved from attempting to falsify a non-falsifiable theory. Popper claimed that, if a theory is falsifiable, then it is scientific.

The demarcation criterion excludes theories that we cannot declare scientific, but it does not say which scientific theories we must keep. For Popper, no methodology rigorously describes the choices made by scientists and he never tried to prescribe such a methodology: he always insisted that the methodology can hardly made rigorous. In particular, a scientific theory that is falsified remains scientific and can still be used in science. This view was in contradistinction to the claimed rigor of the inductive methodology of the logical positivists.

This one also under the subtitle "Demarcation and growth of knowledge":

In contradistinction with the verificationists, Popper never meant that science grows by applying fixed logical rules to scientific statements or theories. In Popper's view, unfalsifiable systems such as logic, mathematics and metaphysics, which have not themselves emerged from fixed logical rules only, play a role in the growth of science. He wrote that important metaphysical research programmes have influenced the development of physics since the days of Pythagoras and Heraclitus. Thornton explains that science, like virtually every other human, and indeed organic, activity, Popper believes, consists largely of problem-solving. Science, in Popper's view, starts with problems rather than with observations. Problems emerge from internal conflicts in the theory, between the theory and the metaphysical program or between the theory and observations. In the latter case, these observations do not come first. Thornton writes that, for Popper, "it is in the context of grappling with a problem that the scientist makes observations in the first instance: his observations are selectively designed to test the extent to which a given theory functions as a satisfactory solution to a given problem."

I also taken out this:

To further defend his solution to the demarcation problem against valid criticism, Popper had to establish his critical rationalism on the solid ground of logic. A clean arena for logical falsification was needed. He was aware of issues such as the fact that observations are theory impregnated (see sections #Away_from_dogmatic_falsificationism and #Away_from_naive_falsificationism for this issue and other issues).

Draft of the section

Even if we accept that angels exist, the sentence "All angels have large wings" is not falsifiable, because though we know how to observe the absence of large wings, no technology (independent of the presence of wings) exists to identify these angels.

In Popper's view of science, statements of observation can be constructed within a logical structure independently of any factual observations. The set of all purely logical observations that can be created within the logical structure constitute the empirical "basis". Popper calls them the basic statements. Popper insists that a basic statement does not have to be possible in practice. It is sufficient that the values or properties that appear in the observations are associated with technologies. Popper gives the example of an apple that moves from the ground up to a branch and then starts to dance from one branch to another. It is a valid potential falsifier for Newton's theory, because we can measure the position of the apple at different times. In contrast, "all men are mortal" is a classical example of an unfalsifiable statement, because no technology can decide that a man is immortal. Yet another example of an unfalsifiable statement is "all angels have large wings". It is not falsifiable, because though we know how to observe the absence of large wings, no technology (independent of the presence of wings) exists to identify these angels.

In more than twelve pages of The Logic of Scientific Discovery (Popper 1959, sec. 14-15, 28), Popper discusses informally which statements among those that can be constructed in the logical structure are basic statements. The logical structure has universal classes. For example, the concept of swans define a universal class. It corresponds to a set of properties that every swan must have. It is not restricted to the swans that exist or will exist. Once we know the universal classes, informally, a basic statement is simply a statement that concerns only a finite number of specific instances in these classes. In particular, an existential statement is not a basic statement, because it is not specific about the instance. For example, the statement "there exists a black swan" is not a basic statement. On the other hand, "this swan here is black" is a basic statement. Popper says that it is a singular existential statement. So, basic statements are singular existential statements. Thornton says that the basic statements are statements that correspond to particular "observation-reports". He then gives Popper's definition of falsifiability:

"A theory is scientific if and only if it divides the class of basic statements into the following two non-empty sub-classes: (a) the class of all those basic statements with which it is inconsistent, or which it prohibits—this is the class of its potential falsifiers (i.e., those statements which, if true, falsify the whole theory), and (b) the class of those basic statements with which it is consistent, or which it permits (i.e., those statements which, if true, corroborate it, or bear it out)."

— Thornton, Stephen, Thornton 2016, at the end of section 3

As in the case of actual falsifiers, decisions must be taken by scientists to accept a logical structure and its associated empirical basis, but these are usually part of a background knowledge that scientists have in common and, often, no discussion is even necessary. However, this agreement, if we can speak of agreement when there is not even a discussion, exists only in principles. When an actual falsifier is proposed, the technology used is then considered in details and, as we discussed before, an actual agreement is needed. This may require that we use a deeper empirical basis, hidden within the current empirical basis, to make sure that the properties or values used in the falsifier were obtained correctly (Andersson 2016 gives some examples).

Dominic Mayers (talk) 19:35, 12 April 2020 (UTC)

Reusing texts that was taken out

Now, it's all done. I implemented 1 to 8. It remains to reinsert some of what was taken out. The only texts that I feel need to be considered for reinsertion are these ones:

The demarcation criterion excludes theories that we cannot declare scientific, but it does not say which scientific theories we must keep. For Popper, no methodology rigorously describes the choices made by scientists and he never tried to prescribe such a methodology: he always insisted that the methodology can hardly made rigorous. In particular, a scientific theory that is falsified remains scientific and can still be used in science. This view was in contradistinction to the claimed rigor of the inductive methodology of the logical positivists.

This one also under the subtitle "Demarcation and growth of knowledge":

In contradistinction with the verificationists, Popper never meant that science grows by applying fixed logical rules to scientific statements or theories. In Popper's view, unfalsifiable systems such as logic, mathematics and metaphysics, which have not themselves emerged from fixed logical rules only, play a role in the growth of science. He wrote that important metaphysical research programmes have influenced the development of physics since the days of Pythagoras and Heraclitus. Thornton explains that science, like virtually every other human, and indeed organic, activity, Popper believes, consists largely of problem-solving. Science, in Popper's view, starts with problems rather than with observations. Problems emerge from internal conflicts in the theory, between the theory and the metaphysical program or between the theory and observations. In the latter case, these observations do not come first. Thornton writes that, for Popper, "it is in the context of grappling with a problem that the scientist makes observations in the first instance: his observations are selectively designed to test the extent to which a given theory functions as a satisfactory solution to a given problem."

The subject is already covered in subsection falsifiability#Away_from_naive_falsificationism, but it so important that it needs to be reconsidered in the light of the definition. This issue together with the problem of the empirical basis, are at the source of almost all controversies. In both cases, the issue is that Popper, one might feel, hides a lot under the part that, he says, cannot be made rigorous. Therefore, it's very natural to question whether its true that only logical deductions and falsifications can be made rigorous in the scientific process. Is Popper correct when he says that the induction problem is solved, because the growth of science, except for the part where we do logical deductions and falsifications, is not a rational activity, but only an organic activity, a natural process of evolution? Is he correct that no induction is responsible for progress of science in the same way that human physiology was not created by our mental process? One might think that I am mistaken and that Popper only meant that it is a psychological process that does not concern epistemology. But, if this was the situation, he would have agreed with Hume, because that was Hume position. It must be said, though, that what Popper said is not that it is not a psychological activity, but only that it was not a logical activity at the psychological level. As soon as we accept activity that cannot be rationally described in terms of some logic at the psychological level, i.e., an activity that could be described as a viewpoint on what he prefers to call an organic activity, his argument to reject (what was called at the time) psychologism does not apply. Dominic Mayers (talk) 02:02, 14 April 2020 (UTC)

I changed my mind. It's an interesting subject, but I think the article does not have to go into this. At this stage, it's more important to consider how falsifiability works in the context of probabilistic laws and how it relates to the principle of parsimony. Dominic Mayers (talk) 15:04, 14 April 2020 (UTC)

---

Problem with The demarcation problem section

Basically, until a recent modification by user:Biogeographist, the section presented clearly Popper's position in contrast with a previous understanding. After the modification, it only appears as a difference of view point between Popper and a few other philosophers, which is totally not interesting. There is something wrong that, after this modification, we are not saying anymore that Popper really made a clear contribution. The key point is that to be able to claim that Popper made a clear contribution there should be no need that all others were wrong and that Popper corrected all of them. It sufficient that the knowledge was clearly not there in the discussions among others. For example, what the article [2] does is useless for our purpose. What it does is that it extracts from many divergent views, none of them being close to Popper's view, a minimal consensus that can be seen as consistent with Popper's view. But, this minimal consensus is not elaborated at all. It's very vague and it's not the view of any philosopher at the time. In contrast, Popper's view is very elaborated and detailed and it's not an imaginary constructed view. I need to think about that, but it's wrong that we cannot find a way to say clearly that Popper answered an important open question at the time and cleared out a lot of misunderstandings. Well, whether it was a solution and the other views were misunderstandings is not something Wikipedia can express, but we can contrast Popper's view with these other views.Dominic Mayers (talk) 04:32, 19 April 2020 (UTC)

I think the solution is to take separately the main views that were considered significant at the time, not a minimal consensus. What's is the point of a minimal consensus in science or in philosophy of science? What a strange idea? In contrast, if only one of the main views, already provided the solution that Popper provided, then Popper did not made an important contribution. Otherwise, he did and we can see it by comparing it with these views. Dominic Mayers (talk) 04:52, 19 April 2020 (UTC)

Dominic Mayers wrote: the section presented clearly Popper's position in contrast with a previous understanding. After the modification, it only appears as a difference of view point between Popper and a few other philosophers, which is totally not interesting. And: In contrast, if only one of the main views, already provided the solution that Popper provided, then Popper did not make an important contribution. But it is just your opinion that the details of controversies among particular philosophers are totally not interesting! The past is what it is, whether you find it interesting or not. In fact, scholars have found such controversies among particular philosophers to be interesting enough to write whole monographs about them. (The book series Vienna Circle Institute Yearbook comes to mind as addressing debates among Vienna Circle philosophers, among other issues.) And the idea that Popper did not make an important contribution if this aspect of his thinking was shared by others also strikes me as just your opinion. Looking at the issue in terms of later influence, Popper's contribution has to be considered important regardless of whether other philosophers shared this aspect of his thinking. Don't forget that the Vienna Circle is not the only context that should be considered (that is, the whole field of philosophy of science was not centered in Vienna!) but also, as mentioned above, what Niiniluoto and Laudan called the nineteeth-century methodological tradition, including Whewell in England and Peirce in America, and the other thinkers that Lakatos mentioned in "Falsification and the methodology of scientific research programmes". Biogeographist (talk) 11:23, 19 April 2020 (UTC)

Biogeographist, you mention references that refer to disagreement between philosophers. Of course, this happens. For example, it could happen simply to show that there were different views within a group. This can be interesting, but it's not at all the point that is being made in the section. So, it remains that, even though the general context that you present is interesting, a particular disagreement between Popper and two or three other philosophers is not interesting in itself. This particular point within the section as it is now is most likely not verifiable (in the sense that it would be cherry picking). It is not a point that deserves to be mentioned in itself. It has to appear within a context that uses it to make another point, which other point would then be verifiable. On the other hand, I agree that finding out a minimal consensus among a large group of philosophers is useful. I was wrong when I wrote that it is a strange idea. In any case, the section does not make use of that minimal consensus. Dominic Mayers (talk) 12:42, 19 April 2020 (UTC)

The view that should be made in the section, which is certainly verifiable, is simply that the demarcation problem was central in the Vienna circle and that the solutions provided all tried to use observations to verify statements. This is consistent with another view, which is likely to be also true, which is that none of the solutions obtained consensus. Dominic Mayers (talk) 15:51, 19 April 2020 (UTC)