(Im going to make some references to Critical Rationalism, since I don't know where the line is drawn between CR and CF).
Arguments Should Be Decisive Criticisms
Decisive criticism would be great,if available. You said that proof is difficult because you have to show that a theory works in every case. But there can be criticism of criticism, counter counter argument against counter argument. (Even criticisms of critical rationalism!). So, to know that a criticism is decisive, you have to know that no one could possibly come up with a counter criticism. (And that no one could reframe the whole arena). Which presents a similar problem to the problem of decisive proof. Neither form of perfection is available.
Theoretical questions have a particularly acute version of the problem. In philosophy, even the nature of truth is disputable. Practical decisions where you get instant feedback on success are much easier. The two are not the same.
Decisive positive arguments are either rare or entirely inaccessible
Ditto decisive negative arguments!
Imperfect criticisms and imperfect negative evidence could still weigh more heavily than imperfect support and positive evidence. That still doesn't add up to the CR claims that there is no justification, and induction never works. You are right that induction is dumb, but it still sometimes works..especially if taken probabilistically. The Turkey paradox is an example where it doesn't work, the Sun rising in the East is an example where it does.
Dumb is better than nothing. In an imperfect world, you need to use anything with non-zero value.
A decisive argument (or group of arguments) contradicts the negation of its conclusion, so both can’t be true
Weighting is needed to see which is false. Contradiction alone does not settle anything because it is symmetric. Regarding evidence as weighing more than theory, which is needed to make naive Popperism work, is a form of weighting assumed by default.
I’ve written criticisms of induction and weighted factors before.
Weighting isn't adding apples and oranges , it's adding value_of(n apples) and value_of(m oranges). Everything gets converted to the same type first.
Yes, it's indecisive, but you have to use weighting , so you are stuck with it.
CF’s most important idea involves distinguishing between decisive and indecisive arguments. CF evaluates ideas by whether they succeed or fail at a purpose (achieve a goal, solve a problem, answer a question). Ideas are for something and can’t be evaluated in isolation
That's two ideas, and they are somewhat at odds.
Something can seem decisive in a limited context, but not in a wider one. Since we do not have perfect knowledge of the widest possible context, we can't say acontextually that anything is decisive. And we can't "just see" that something is "just true". Deep theoretical issues are different to shallow empirical ones.
All our arguments are fallible
Then none are decisive!
A single counter-example or error can be a decisive criticism without considering vast numbers of possibilities. Popper discussed this with examples like “all ravens are black”, which is refutable with one contradictory observation, but still not proven with a million compatible observations.
Well .. it's not as simple as naive CR makes out. A single observation can be erroneous (eg Martian canals, cold fusion). Hence the norm of replication in science. Also , theories can be patched to handle objections (Kuhn, Quine, Duheim). Etc.
If indecisive arguments are logically flawed, then why do they often appear to work moderately well
Indecisive arguments dont have to be logically flawed...they can be reframed as valid probabilistic arguments. So long as you are willing to abandon certainty...which you are if you are a falliblist!
Positive arguments point out good trait(s)
Positive evidence...is evidence, not argument
CF’s main motivation is logical arguments showing that various other approaches cannot possibly work.
I don't think you have shown that. You haven't shown that decisive falsification always works, or that the alternatives never do.
Thanks for engaging.
So, to know that a criticism is decisive, you have to know that no one could possibly come up with a counter criticism.
I think you didn't take into account the definition that I used: "A decisive argument (or group of arguments) contradicts the negation of its conclusion, so both can't be true." Excluding the possibility of counter criticism is unnecessary for this definition to be met. The point is that if A and B could both be true – if they're compatible – then it's problematic to view B as a criticism of A.
Neither form of perfection is available.
The goal is basically logical relevance, not perfection.
You are right that induction is dumb, but it still sometimes works..especially if taken probabilistically.
For induction to work, it'd have to define steps a person can follow to induce a theory. It'd have to specify what constitutes inducing a theory. The main issue with induction isn't the quality of the results, but actually defining a specific method that produces any results. Over the years, I've never been able to get an answer to this along with a worked example and answers to basic questions like which of the infinitely many patterns fitting the data should be induced and which shouldn't and why those.
Weighting is needed to see which is false.
When two things contradict and you're deciding what side to take, weighting them and choosing the higher weighted side is one approach. But it's certainly not the only approach. Since you're just choosing between two things, quantitative evaluation seems less relevant or appealing than in many other scenarios.
I could go into more detail here and it's an interesting topic but I think I've written enough for an initial reply so I'll leave it at saying I don't see what aspect of contradiction-resolution makes quantitative approaches mandatory. My best guess is you think they're always mandatory for everything, which might be better approached from another angle, not via this sub-problem.
Weighting isn't adding apples and oranges , it's adding value_of(n apples) and value_of(m oranges). Everything gets converted to the same type first.
My link discusses dimension conversion (like from apples to value) being problematic. That's covered.
All our arguments are fallible
Then none are decisive!
Do you think fallibilism prohibits reaching conclusions? Decisive basically means conclusive, aka adequate to tentatively, fallibly reach a conclusion, as against arguments that don't provide that much (where accepting the truth of the argument, as a premise, would still be inadequate to reach a conclusion).
Well .. it's not as simple as naive CR makes out. A single observation can be erroneous (eg Martian canals, cold fusion).
Popper knew that and wrote about it.
Indecisive arguments dont have to be logically flawed...they can be reframed as valid probabilistic arguments.
Do you have an example? If it's actually valid, I might tell you it's decisive. As above, decisive is an easier standard than you interpreted it as. I'm not sure what sort of probabilistic argument you have in mind though.
Epistemic learned helplessness should control here, even if it's inconvenient for rationalists trying to argue people into unusual beliefs.
What do you mean?
Here's my best guess, but this is low confidence. I presented a non-mainstream view of epistemology. You are not an expert on epistemology. So you will defer to people you see as experts on this topic without engaging with my arguments (like your link talks about using history as an example). If that's what you mean, that's fine, but I think this site is a reasonable place to find people to engage with about epistemology.
The point is that at least in most cases decisive arguments shouldn't work, and if said to rational people, won't work, because of epistemic learned helplessness.
Isn't your point about all arguments, not just decisive arguments? What does it have to do with my discussion of which types of arguments are logically and epistemically better than other types of arguments?
The idea is implicit that you should use decisive arguments because then people will have good reason to believe them.
Are you trying to say we should use worse forms of argument on purpose because of epistemic learned helplessness? I don't see how that would help and you haven't given any analysis about that. Epistemic learned helplessness is a separate issue from what I was talking about: when using arguments, which types are impersonally best, just looking at the subject matter and arguments themselves? I wasn't talking about human behavior or psychology.
I wasn’t talking about human behavior or psychology.
If you do not intend to get humans to believe the arguments, you are correct that epistemic learned helplessness doesn't apply. I do find this sort of odd, however.
Critical Fallibilism (CF) is a philosophy I developed which deals with rationality, knowledge, and critical thinking and discussion. It builds most on Karl Popper's Critical Rationalism, which says we learn (create knowledge, solve problems) by an evolutionary process of conjectures and refutations. Popper rejected positive arguments (justifications) and induction. He advocated fallibilism and error correction.
CF's most important idea involves distinguishing between decisive and indecisive arguments. CF evaluates ideas by whether they succeed or fail at a purpose (achieve a goal, solve a problem, answer a question). Ideas are for something and can't be evaluated in isolation. The same idea can succeed at one purpose and fail at another (and indeed all ideas do that). Negative arguments (criticisms) argue that one or more ideas fail at one or more purposes. A purpose could be avoiding dehydration, having true ideas, or both. Any group of purposes joined with "and" (or other logical operators like "or") is a purpose.
A decisive argument (or group of arguments) contradicts the negation of its conclusion, so both can't be true. A decisive positive argument contradicts its target being false (it says the idea must succeed at its purpose). A decisive negative argument contradicts its target being true (it says the idea must fail at its purpose). If you accept the argument, you shouldn't accept the thing it contradicts, because they're incompatible.
Decisive negative arguments are reasonably common and accessible. We can find and point out errors such as counter examples or flawed logic. One error is often enough to fail at a purpose.
Decisive positive arguments are either rare or entirely inaccessible. Pointing out 1000 good things isn't enough to prove an idea will succeed at its purpose. Despite all those merits, there could still be an error that causes failure. Basically, errors have logical priority over positive traits. No matter how many good traits an airplane has, a single mistake in the engine can cause a crash.
Karl Popper argued that negative arguments are better than positive arguments. While he made several good points, I think there's another issue that Popper missed. Decisive arguments are better than indecisive arguments, and most or all of our decisive arguments are negative. Why? Decisive positive arguments require 100% proof; otherwise you could accept the argument, and accept that its conclusion is false, without contradiction. I think fallibilism excludes 100% guarantees against error in math, logic and every field. Even if you disagree about how broad fallibility is, you could still accept CF for science, philosophy and most fields.
Criticism is easier than proof because proof requires addressing all possibilities. A single counter-example or error can be a decisive criticism without considering vast numbers of possibilities. Popper discussed this with examples like "all ravens are black", which is refutable with one contradictory observation, but still not proven with a million compatible observations.
All our arguments are fallible and can be reconsidered with new ideas and evidence. But if you accept an observation of a white raven and some background knowledge, that contradicts "all ravens are black". There's no comparable way to prove it, even fallibly.
You can make decisive positive arguments using universal premises (e.g., "All men are mortal.") which assert their own completeness. But that just moves the problem: how do you prove that premise since observing a million mortal men is inadequate?
Indecisive negative arguments have the same basic flaw as indecisive positive arguments: you can accept the argument but still reject the argument's conclusion without contradicting yourself. There's no logical connection between the argument and the conclusion. Logically, indecisive arguments don't do anything. Contradiction is a powerful, useful logical tool used with decisive arguments, but we don't have good alternative tools to use with indecisive arguments. Compatibility (non-contradiction) lacks logical power despite sometimes misleadingly being called "support" or "confirmation".
If I observe a purple hamster or red tree, that is compatible with "all ravens are black"; compatibility is basically worthless without also using another concept like relevance. But even with a million relevant, compatible observations – e.g., observations of black ravens – a white raven could still exist. Relevance is very difficult to define and evaluate, but even if we had perfect knowledge of what was relevant, compatibility plus relevance would still be inadequate to make indecisive arguments effective. Nothing about seeing a lot of black ravens actually means that there couldn't be a white one.
If indecisive arguments are logically flawed, then why do they often appear to work moderately well? Many of them are convertible to decisive arguments. They have a valid point which is presented imperfectly. Similarly, many positive arguments are convertible to negative arguments. I hypothesize that arguments which can't be converted to decisive, negative form are wrong. Skipping the conversion step is often reasonable in low-stakes, low-precision, friendly contexts.
How do you convert arguments? Positive arguments point out good trait(s). To convert to negative form, criticize alternatives for lacking the good trait(s). Indecisive arguments involve more creativity to convert. They point out good or bad things without logically connecting that to success or failure at a purpose. To make them decisive, clarify the purpose, figure out criteria for success and failure, and point out how the bad things (or missing good things for alternatives) cause failure.
Focusing on negative arguments can take some getting used to because it's more indirect. Instead of arguing in favor of an idea, you criticize alternatives to that idea. You also try to criticize the idea. The conclusion you reach is the idea you can't find any decisive error in, despite trying.
If you have multiple ideas that you think will work, you can use any of them, or you can aim for a more ambitious purpose/goal. CF has methods for narrowing it down to exactly one non-refuted idea, but that's often not worth the effort. If you have no ideas you think will work, you can brainstorm/research more, adjust your purpose/goal to be easier, or give up and do something else.
If you don't find an error with an idea, it could still be wrong. Why use it? Because ideas you don't know are wrong are preferable to ideas you do know are wrong (already found a decisive error in). It never makes sense to use an idea for a purpose if your best understanding is that it will fail at that purpose.
Why?
What is the motivation for CF? A reasonable first impression would be that CF doesn't sound strictly wrong (you don't see a reason it can't possibly work), but it sounds inconvenient or cumbersome. So what's the upside to make it worth pursuing?
CF's main motivation is logical arguments showing that various other approaches cannot possibly work. CF prioritizes correctness over everything else. (I actually think CF is reasonably convenient and elegant once you get used to it, and has various merits besides correctness, but I acknowledge there's an initial learning curve.)
Two main alternatives to CF are induction and weighted factors. Serious problems with these are actually well known in the academic literature and have been written about by advocates of these approaches, not just by opponents like Popper. I'm not actually saying anything very new by claiming these approaches are flawed. I think people try to use them despite the known errors because they've largely given up on finding alternatives, and errors seem somewhat ignorable to people accustomed to indecisive thinking. Some people are ignorant of the problems with their approaches, but I think many experts are instead pessimistic about finding something better. Ignorance of problems reduces motivation to consider alternatives. And experts who understand the problems may not want to spend time studying or debating a new system that they find initially counterintuitive and aren't optimistic about.
The various schools of inductive thought can be seen as attempts to say that indecisive arguments sometimes partially work. Compensatory weighted factor approaches (where a high score at one factor can compensate for low scores at other factors) also try to use multiple indecisive arguments to reach a conclusion. Non-compensatory multi-factor approaches are less common but exist and have some overlap with CF.
I've written criticisms of induction and weighted factors before. For this article, I'll just say that I'm open to discussion if someone disagrees but wants to try to reach a conclusion about the matter. In a discussion, I only have to address one person's concerns instead of all potential concerns, so it can be easier and more focused (and provide valuable feedback and potentially criticism for me, too). I can provide a free account on my forum if someone accepts this invitation (email me).
Examples
The idea "go to McDonald's for a burger" succeeds at the purpose "get lunch" but fails at the purposes "get a gluten-free lunch" and "learn knitting". We should evaluate {idea, purpose} pairs rather than ideas alone. Being more precise, we can include context and evaluate a triple: {idea, purpose, context}.
Note: In other articles I've said "goal" instead of "purpose" and abbreviated the triple to IGC, but they mean the same thing. It's {idea/solution/answer/plan/explanation/argument/reasoning/option/alternative, purpose/goal/question/problem/objective/criterion, context/problem-situation/background-knowledge/scenario} triples. The triple is meant to work with any of the terms separated by slashes; you can use whichever is most natural. Sometimes answer and question work well, but in other contexts solution and problem make more sense. {idea, purpose/goal, context} is particularly generic and works well when you want to keep terminology consistent. Plurals like "goals" or "criteria" may be used because the goal is often a group: the goal could consist of multiple conjoined sub-goals or be to meet multiple criteria. Interestingly, goals and contexts are types of ideas which can be put in the first spot of their own IGC triples. Similarly, arguments are used to evaluate IGCs but they are also part of their own IGCs: arguments have goals and contexts and an argument's IGC can itself be criticized.
A criticism can apply to multiple {idea, goal} pairs. If my goal is to get a gluten-free lunch, the criticism "buns have gluten" applies to McDonald's and also to other ideas like Burger King and Wendy's. The same criticism can also work for other goals, e.g., getting a gluten free breakfast or dinner. It's often important to think using principles and make your criticisms broad enough to cover whole categories of ideas and goals instead of just an individual idea-goal pair.
An indecisive positive argument is "I'll go to McDonald's because their food tastes good". It's indecisive because the reasoning (McDonald's food tastes good) doesn't contradict "I won't go to McDonald's". Both can be true. Also, alternatives like "I should go to Burger King" aren't contradicted even though they imply not going to McDonald's (in a context where you're picking one restaurant for a specific meal). Indecisive means there's no logical problem with accepting the reasoning but also accepting a contradictory conclusion.
To convert to a negative argument, figure out what positive trait is being praised (food taste) and criticize alternatives for lacking it. E.g., "I won't go to Burger King because their food tastes bad." After converting, always check if the result is true. If you actually like the taste of Burger King's food, then it's wrong. In that case, the original indecisive, positive argument was also wrong to conclude that you should go to McDonald's specifically (not Burger King) because of taste.
To convert to a decisive argument, you often need to clarify your purpose/goal. Let's say I'm hungry now and I want to get restaurant food within five minutes of walking. Then an indecisive argument is "I'll go to McDonald's because it's close." A decisive argument is "I won't go to Burger King because it's over a five-minute walk away." More generically, "I won't go to any restaurant that's over a five-minute walk away." can rule out the idea of going to Burger King, Wendy's and more with a single criticism. To make location decisive, we figure out what proximity constitutes success or failure and we criticize ideas that fail.
Next Steps
If CF interests you, a good place to start is by categorizing arguments you make, read or hear as decisive or indecisive and positive or negative. You can do that without changing what arguments you use or like; it can start as a research project for informational purposes. A later step could be trying to convert some arguments to be decisive and negative and paying attention to how some convert well but others don't.
A more advanced issue is considering whether enough decisive, negative arguments exist (including via conversion) to do all of our reasoning with them, or if we need to use some non-convertible positive and/or indecisive arguments. Many people like decisive, negative arguments and see them as highly powerful and useful, but believe they're too scarce to use exclusively. A related issue to ponder is: Can you ask a series of yes or no questions to explore any issue, or are other types of questions (or non-questions) absolutely necessary? Yes or no questions are good at exposing decisive issues. "Does it cost under $50?" invites a decisive answer while "How expensive is it?" or "How happy am I with the price?" don't.