Critical Fallibilism (CF) says to evaluate ideas as refuted or non-refuted using decisive, critical arguments. And CF says to evaluate ideas in the context of {idea, goal, context} (IGC) triples, so the same idea can have multiple evaluations. CF follows Critical Rationalism (CR) in denying the effectiveness of positive (justifying) arguments or induction. CF agrees with CR about accepting fallibilism and denying the possibility of absolute proof. CF adds that arguments don't have strengths: either they refute an IGC or they don't refute it. There's no partial refutation; indecisive arguments simply don't refute IGCs; arguments that can't decisively refute anything don't get partial credit.
I have often been thinking in conventional CF-contradicting terms and found them useful in some way. I then try to think through why I believe the CF ideas are true and what the nature of the contradiction is like.
Overall the article addresses concerns I’ve had over this issue. I think knowing about the conversion patterns is really helpful. The “mostly true/false” case is something I’ve been thinking about. I think I’ve had similar ideas as to what you describe here (but not very clearly) and I like your solution.
When we evaluate what percentage of IGCs are non-refuted, it’s of the selection of IGCs we’re looking at, which we’ve chosen for relevance, and which hopefully are a representative sample of all relevant IGCs.
Should we think of the IGCs as kind of “atomic” for this? I’m thinking that as opposed to chaining a bunch of IGCs together with AND gates. Or keeping the idea and/or context constant and making a goal that can be viewed as a bunch of different goals in an AND chain. In the AND chain case you’re aiming for non-refutation, but in the a bunch of small IGCs (or just a bunch of different goals) case you’re trying to get a high percentage.
In general, I don’t purposefully view getting a high percentage of true IGCs as a goal. I focus on figuring out at least one specific IGC I think is true that I want to act on.
I don’t know if this is what you had in mind or not but FYI comparing ideas by one gets 70% true IGCs and one gets 65%, so the first is better, is just the sort of indecisive, weighted thinking that CF rejects (even though it’s phrased in terms of IGCs, it’s still basically the same thing as doing that with credences).
No, that’s not what I had in mind. I don’t think CF advocates that at all. I was talking about the “mostly true” cases. I see now that’s not at all obvious from my post, because I was holding more context in my head than I wrote out.
I was thinking that for the non-refutation CF advocates you can view complex goals as often being lots of simpler goals chained together with AND gates, and you want the idea to work for every one of the simpler goals, hence the AND gates. But when things contradict CF but are “mostly true”, it’s better viewed as many IGCs with small and simple goals. And “mostly true” means a high percentage of them succeed, but at least some IGCs are wrong so we can’t AND them together.
In general, when a positive argument says “A is good because it has trait B”, it can be reframed negatively as “Alternatives to A which lack trait B are bad”. This is more precise for various reasons that CR discussed and because some alternatives also have trait B. This argument only helps A against some alternatives. Or in other words, it’s a criticism of a category of alternatives but it doesn’t criticize some other alternatives at all. Viewing it as a positive argument for A obscures that it’s actually a criticism of B but not a criticism of C: it does nothing to make A better than C.
We can make this more precise again. The real issue is “A is good because it has trait B which addresses problem P”. The negative version is “Alternatives to A which lack any solution to P are bad.” An alternative with trait B is fine and an alternative with some other solution to P (to the same problem that B addresses) is also fine. You could also list multiple separate problems that B addresses instead of treating it as one problem, but the number of problems isn’t important (it depends on how you conceptually organize and count the problems rather than having any fundamental importance).
In the first of these two paragraphs you refer to a “trait B” and also to a “B” (as opposed to A and C i.e., an alternative idea?). Am I right that these two are different things?
Then, in the second paragraph you refer to trait B as “trait B” and just “B”.
Either I’m misunderstanding what you’re saying or there is mistake with the symbols used. The “B” in the second paragraph seems to mean “trait B” but in the first paragraph “B” seems to mean something distinct from trait B.
Another terminology translation is “I’m not sure why you think that” → “I’m not clear on why you think that”. You can express a lack of understanding or lack of clarity rather than a lack of sureness or certainty. I don’t think it’s a big deal and I often use the more common, natural phrasing even when the rephrasing is barely longer and is easy to understand. I do think it’s worth noticing the issue and changing for some contexts. I don’t think I should use “sure” like that in a CF essay.