Talk:Falsifiability

Untitled

If someone would please tell me where my language is so complex, I would be happy to change it.

the_photon ----- (sorry, no tilde on my keyboard) — Preceding unsigned comment added by the_photon (talk • contribs)

I have removed the complex tag; I have also rewritten (and moved the article).

What is still needed:

• We need a much better (more scientific example)
• Criticisms (look at the EnWP page)

Other improvements welcome as well...--Eptalon (talk) 10:17, 21 December 2008 (UTC)

This article ignores the most important aspect of falsifiability

Popper has always insisted that falsifiability is a formal or logical demarcation criterion. Once the background knowledge is fixed, falsifiability can be logically defined in a non ambiguous manner. He was aware that the background knowledge is never really fixed, but he argues that scientists can agree on such a background knowkedge and, if they cannot, no discussion is possible anyway. He was aware that this criterion required more than a background knowledge - it also had to be complemented with a method to actually select a better theory, but he felt that this other part could hardly be made rigorous. The ingenuity of Popper was to have separated what could be said logically from what could not. This separation is fundamental. Describing falsifiability without this separation is a complete misinterpretation of Popper's work. Dominic Mayers (talk) 13:10, 22 June 2018 (UTC)

Just to add that the method could perhaps be formalized, but the problem would remain that whether it leads to a better theory cannot be proven. As pointed out by some, as soon as we try to provide a logical method to select a theory based on singular observations, we are back to the original problem of induction. It is unfair to say that Popper did not really solve the problem of induction because of that. He called the set of methods used by scientists metaphysical research programs. The term "metaphysical" shows that he was aware that it had to be separated from the logical part. Yes, we can fairly say that a fundamental question is raised: how these metaphysical programs can succeed?

If you think you can put t in simple terms, then go ahead and change the article. --Eptalon (talk) 13:55, 22 June 2018 (UTC)

Actually, my most important concern is this tagline "the possibility of a statement to be proven wrong". The problem is that "falsifiable" does not mean "possible to be proven wrong". It's both too weak and too strong. It's too strong, because nothing empirical can be proven wrong. One can always argue that the experiment went wrong. This is well understood by all philosophers, not only Popper. Therefore, for that reason it is certainly not the definition of falsifiability adopted by philosophers. It is also too weak. It is too weak, because even the statement 2 + 2 = 5 is falsifiable, if we use this definition. Remember that this is a demarcation criterion for scientific statements. Is 2 + 2 = 5 a reasonable scientific statement? No. Therefore, for this reason also, it is not uthe definition of falsifiability.

I would not be that concern, if this tagline would not be used as a tagline even for the (ordinary) wikipedia Falsifiability article when we see it on a mobile phone. This pollutes even the standard Wikipedia article.

There is only one correct way to define falsifiability (as introduced by Popper) and it requires that we explain a simple thing before the definition can be given. We need to understand that some scientists rarely go and do an experiment. Instead, they refer to simple statements that have an interpretation in terms of experiments. The point here is that a theory is first expressed in a purely formal manner, using statements. It's the relation between the statements that define the theory. The fact that some of these statements have an interpretation in terms of experiments is important, but to define falsifiability, the details of this interpretation is not important. Falsifiability is a quality that is given in terms of the statements and their relationship. This makes it a much weaker criterion. Someone cannot argue anymore that the experiment went wrong, because the definition does not care about that. It is sufficient that the theory is contradicted by a basic statement that is known to be interpretable in terms of an experiment - the details of an eventual experiment do not matter. It is also stronger, because we also require that the statement is consistent. So 2 + 2 = 5 is not falsifiable, rigorously speaking, because, before it can be falsifiable, it must be consistent. Dominic Mayers (talk) 01:31, 23 June 2018 (UTC)

┌─────────────────────────────────┘
I removed the example of the 1915-ish experiment, and added a simpler one (All Swans in Europe are White, so until a Dutch explorer found a black one in Australia, Europeans believed Swans are white). I also added a paragraph about how to actually falsify a theory. Personally, I would not cgo with numbers and Algebra, because they have a myriad of philiosophical problems of their own. I don't know, but we probably need to mention Kant/Fichte/Schopenhauer when we talk about obeservations. We might also mention Russel's "On Denoting" (The current king of France...), and finally Hempel with his Raven . I miht be wrong, but in my opinion, a priori knowledge cannot be falsified. Please go ahead, and change the article as you see fit. If the article in itself is correct, we can later try to simplify the language to fit SEWP. --Eptalon (talk) 07:48, 23 June 20y18 (UTC)

It's a very good idea to put falsifiability in the context of Kant, Russell, etc. However, my point is that the first step is to well define falsifiability. Russell, Kant, etc. did not have their own definition of falsifiability. To my knowledge, as far as I can see in the literature, falsifiability means Popper's falsifiability.

Yes, Russell and Frege debate on the question of singular vs universal concept or name is interesting and could help to put falsifiability in its context.

Similarly, the notion of a priori knowledge of Kant, his solution to the problem of induction, is of course very much related to the notion of falsifiability.

However, if we go into this without having explained what is falsifiability we will only create confusion.

It's not complicated. A child could understand that, but we first need to explain that theories exist on papers, by which I mean that they are statements. Perhaps a good analogy is a drawing that represents some scenery or the plan of an architect that represents a building. These plans are constructed following some rules. They have their own rules or structure. In the same way, a theory is made of statements and exist separately from the physical world. This is not that complicated.

Plans have properties. You can say that a plan of a house has no door, for example. You don't need to see a real house to determine that the plan has no door, and thus is not a very good plan. In the same way, falsifiability is a property of a theory and you don't need to consider any interpretation of the theory in terms of observation to determine if it is falsifiable or not.

This is a very simple thing, but it is very important that it is clear before we go and discuss Russell, etc. --Dominic Mayers (talk) 19:10, 22 June 2018 (UTC)

I have again changed the article a little. It woul be good if you could adapt it ot what you think it should look like, we can then work from there.--Eptalon (talk) 12:31, 23 June 2018 (UTC)

The article, as currently written, tells the story of the wrong (inductive) reasoning of the Europeans that concluded that all swans are white and the subsequent expedition that discovered the black swans (for the Europeans). The story works well to explain the problem of induction. Here is another one about turkeys, instead of black swans, but the angle is different, a bit more intense. These stories work well to explain the problem of induction. They also give the idea why, though we cannot prove by experiments that a theory is true, we can prove that it is false. As these stories illustrate, as far as logic is concerned, only one counter example, show that the theory is false.

However, falsifiability was introduced by Popper to address a much more delicate issue, which is the problem of the empirical basis. To explain the problem, I will continue the story of the expedition - it is not the true story, but it conveys the idea. The expedition came back and told about the black swans. No one believed them. They said that they were only dirty white swans, the crew members were drunk, etc. So, they organized another expedition. Again, the people did not believe them. They said that they were paid by the first expedition, etc. Yet another expedition was organized. This time they believed them. They thought they must really have seen black swans, but perhaps the swans were sick and they will die soon and the theory would still be true. They organized several expeditions and each time, the expedition reported having seen black swans on the black swan river. Using the principle of induction, they concluded that every time that we look at this specific area we see black swans. But, Popper and other philosophers came in and said, no, we are trying to show that the principle of induction is invalid by proving that not all swans are white. So, we cannot use it to conclude that we will always see black swans on the swan river. So, strong believers attached to the white swans theory said that the theory is still valid, because the falsifying statement "there are black swans" was not proven conclusively, arguing that it is a principle of science that a law is true only if it is reproducible. The moral of the story is that there are several issues when we consider the empirical aspect of a falsification.

What was Popper's solution? He said we will only consider the logical aspect of the theory to determine if it is falsifiable. The problems with the empirical aspect will then automatically disappear. So, he wrote on a piece of paper the statement "All swans are white" and he wrote the statement "In 1697, there were black swans on the swan river in Australia". He pointed out that the latter contradicts the former and is something that corroborates the falsifying hypothesis There are black swans. This is a nice property that we don't have with all theories. He called this the falsifiability criterion. Now, when people argued that the statement There are black swans was not proven, he agreed, but he said that does not matter for his criterion. In fact, he always stressed that we must distinguish between falsifiability and falsification. To make it even more clear, he said that there are two meanings to falsifiability. The first meaning is the one in the tagline, it's the common non technical meaning. His definition corresponds to the second meaning: a criterion that is based on the logical form of the theory and considers only the relation between the theory and basic statements such as "In 1697, there were black swans on the swan river in Australia". Dominic Mayers (talk) 19:59, 23 June 2018 (UTC)

The distinction between falsifiability and falsification is misunderstood

If the distinction between "falsifiability" and "falsification" was only that one is the possibility of being falsified and the other was the action of falsifying, would have Popper made a big deal about that? He would not have even mentioned it. Who would have done that? The technical term "falsifiability" in Popper's definition has a different root than in "falsification" - it's not the same meaning of "falsify". This is what he meant, of course. This is obvious, if you understand Popper. In "falsifiability", the "falsify" only refers to the logical form of the theory - it only concerns the relations between the theory and basic statements. So. you can falsify a theory without caring about the actual experiments or observations that are associated with basic statements. Note that Popper never uses "falsify" with that meaning. He reserves that meaning to "falsifiable" and "falsifiability", which, he explains, are technical terms. On the other hand in "falsification", the "falsify" means to actually do these experiments or make these observations. This is the meaning of "falsify" when Popper uses it by itself. In the example of the black swans, the "falsify" in "falsifiability" can be done on papers, but in the other case, for the "falsify" in "falsification", actual expeditions are needed. Note that the "falsify" in "falsifiability" is not enough for a falsification, but it is enough in the meaning of "falsifiability" or "falsifiable". Dominic Mayers (talk) 04:11, 24 June 2018 (UTC)

I understand that this separation between the formal or logical aspect and the actual reality associated with it can be difficult to grasp, especially because the formal or logical aspect by itself, without the interpretation in terms of the real world seems useless. Let me explains this conceptual difficulty with an analogy. I hope it will help. Someone gives you a car, but the car does not have any instruction how to use it, no map of the area to help you use it to go somewhere, etc. It's absolutely true that a car, if you don't know at all how to use it, etc. is totally useless. Is that mean that the person has given you something that is useless? No. If you have the context, you know how to use a car, a map of the area, etc., then it is useful. In the same way, falsifiable theories in accordance with Popper's definition do not mean much. They are falsifiable (in that technical sense), but if you don't have the proper method of science, you would not even try to falsify them correctly, etc. Popper explained that a good scientific method was needed to complement his definition of falsifiability. Unfortunately, some philosophers considered that the scientific method, the rules of the game, where a part of Popper's demarcation criteria. This is perhaps the most important source of confusion about Popper's falsifiability. It's like saying that the maps, the instructions how to use the car, etc. are a part of the car. They are not. It's an important distinction in the case of falsifiability vs falsification. It makes a big difference. Dominic Mayers (talk) 05:05, 24 June 2018 (UTC)

Our readership

The end product to such a discussion needs to be a page which does justice to the concept, but is intelligible to, say, an intelligent native speaker of English who is not a major in philosophy or mathematics. Or maybe a foreign language speaker who does have some background in the topic. Either way, the product has to be written in fairly straightforward English. There are occasions when this simply cannot be achieved, but that is the last thing we want to happen. Macdonald-ross (talk) 10:33, 24 June 2018 (UTC)

I am sorry, if I seem antagonistic. I don't believe that anything can be accomplished, if we are not thinking together. My goal was to help people who work on this article to have a correct understanding of the key concepts behind falsifiability. We can have different opinions on some issues, especially when these differences reflect different views in the literature, books, notable people, etc. But, the issue here is simply the definition(s) of falsifiability. To my knowledge, there is only one technical definition of falsifiability and it's Popper's definition. If there is another, we should give both, but I am not aware of another. In any case, I don't think that not understanding the definition falsifiability when you write an article about it should be an option.

So, the key point is that I discuss here to help people who work on this article to understand the definition of falsifiability. I cannot see how we can work together nicely if we don't have the same definitions, that is, don't use the same language. Dominic Mayers (talk) 01:04, 25 June 2018 (UTC)

Just to say that the fundamental issue still remain.

The article must be easy to understand, but let's not fool ourselves by assuming that the concept of falsifiability is in itself simple. The article as it is now is simply wrong. Simple should not mean wrong.

Please, whoever tries to explain falsifiability in a simple way, you got to understand that it cannot be explained without reference to the fundamental distinction that Popper made between formal logic (which only deals with formal sentences and their relationship) and the method to find, accept, reject, etc theories. This is very fundamental and at the basis of every thing else that Popper explains, including falsifiability. Paradoxically, the "Logic" in "Logic of Scientific Discovery" refers to the method. In order to have a "clean page" to discuss the method (i.e. the logic of scientific discovery), Popper needed to define falsifiability without any reference to a method, in particular, without any reference to an eventual falsification in that method, except that the formal concept of "basic statement" (also called "observation statement" by others) is to be interpreted as an observation in the method. In other words, the only link with the method is the existence of a formal class of statements, the basic statements, which in itself has no meaning at all, but will gain an interpretation as observations in some complex method, which Popper says can hardly be made rigorous.

Do not fool yourself in thinking that falsifiability must have a philosophical or scientific meaning independently of Popper. It does not. Of course, falsification is as old as logic. The Greek knew about falsification, but this is exactly what Popper says should not be confused with falsifiability. The technical terms falsifiability and falsifiable are, obviously, not simply related to falsification as, for example, "provability" and "provable" are related to "proof". If the distinction between falsifiability and falsification was only that, do you think Popper would have made a big deal about that? Of course not. Currently, the distinction between falsifiability and falsification that is explained in the article is just the same as the distinction between provability and proof.

I appreciate that we want to be simple, but we need to explain the correct concept, not a different concept. We must realize that the usual distinction between falsifiability and falsification, though simpler is clearly not a point on which Popper would have insisted. Yet he did insist on a difference between both. So, we must conclude that he insisted on something more fundamental. The current article misses this point, because it emphasises the usual distinction between falsifiability and falsification. Isn't it obviously way too simple and unworthy of mention to be a point emphasised by Popper? Dominic Mayers (talk) 01:53, 16 October 2018 (UTC)

Real world

Much that's said above is obviously sound. However, our readership is likely to be 15-year old schoolkids rather than PG philosophy students. We're sitting in the middle of a population which believes rumours rather than experimental science, where philosophy students are rarer than rare birds. That's what limits our ability to wrote a) full detailed explanations, while b) fulfilling our remit to write for this audience. Macdonald-ross (talk) 07:18, 23 April 2022 (UTC)