Trying to understand justification better

I’m confused re justification so I’m trying to learn to understand it better.

I posted on Twitter at first:

Am I correct to assume: ~justification is a bad idea re whether an idea is true or not, but is valid re eg explaining choosing a particular action? Similar to, as I understand it, probability being bad re good/bad ideas but ok re outcomes? @curi42 @alan_forrester @j_mallone

Curi replied:

use CF forum?

I agree that posting here (CF forum) is a better idea than asking on Twitter. So here I am after being pretty much a lurker ~since commenting on curi.us was removed.

Are you talking about jusitficationism / justified, true belief? The commonly-known meaning of ‘justification’ is more broad than what JTB refers to (and what criticisms of JTB refer to)

Have you read Curiosity – Fallible Justificationism? You could also search for site:curi.us justified and similar.

I think @curi has talked about his ideas on the topic (or a closely related one) recently in some of the DD posts (particularly about the disagreement he had with DD about it).

Edit: you say “~justification”, which makes it sound broad, but I suspect that where you’ve heard it from (or why you’re interested) is more narrow, which is why I ask the first question.

Also, glad you’re here :slight_smile:

I might be. I’m not sure. I don’t think that I have a good understanding of what justified, true belief (JTB) actually is. Checking out JTB closer seems like a useful thing for me to do.

Not yet. Thx for the link.

Thx. I’ll look for that as well. I assume it’s a YT video since you wrote “talked about”. Is that correct?

It’s roughly something like this:
I think that CR and CF say that justifying ideas as true or probably true is a mistake. But from here I am confusing a lot of stuff because I kind of take this to mean that nothing can be justified - which I think is a misunderstanding from my part. Like I do think that you can justify a decision you made (by eg giving an explanation as to why you took that decision - I think that giving an explanation would be to “justify” that decision).
So I am confused about something here, but I do not know what. Hopefully looking into JTB and the link and video (?) that you mentioned will help me out some here.

While writing this I saw in my Yes or No Philosophy folder that it has an article called “Justification and Problems”. I will read that article as well as it’s probably very useful for my problem.

Thank you, @Max. I appreciate that :slight_smile:

I’m not very clear on your goal, like what project this is part of and how justification came up. But the link Max gave is good and Yes or No Philosophy Summary

It’s part of my goal of learning to understand CR and CF better.

I was reading some old LW forum posts of yours (curi) from 2011 and also during these days I was reading posts here on CF forum as well.
I don’t really remember what exactly brought up justification, but it might have been this:

It glorifies him [Aubrey] and he thinks it justifies stuff, like non-intellectual methods, that would otherwise be bad. Ironically he thinks trying to do something so important justifies him worrying about rationality less .

In the LW forum posts I had read eg (page 202):

Regresses all go away if we drop justification. Don’t justify anything, ever. Simple.

I became confused so I asked the Twitter question. (Well, roughly something like this anyway.)

Thx. I started reading those yesterday and will continue to do so these coming days.

The word justify has more than one meaning. Re AdG I was using a standard non-epistemology meaning.

Ok. I think the core to my confusion was only thinking about the epistemology meaning of “justify”.

No, I just meant text discussion in general. curi has made some YT vids that are relevant and might contain things about that, but I can’t remember that particularly.

IMO you should consider the option that you don’t understand the epistemic meaning well enough to judge that. (Like, if you can’t reliably identify the intended meaning of ‘justify’ from context, how could you know what you were focusing on?) Granted, you might have learned things since your first post, but, if that’s the case, well you haven’t discussed those things (which makes it hard to judge).

Edit: btw, I think you’re right, in that you were taking a narrow view of justification

Fair points re not understanding well enough and not discussing.

I think that justification in the everyday sense (non-epistemic) means roughly to give one’s reasons for eg a specific choice or action.
I’m not sure if justification, in this sense, can be used both when one is giving sincere reasons as well as when one is lying to oneself/others. I think it can be used in both sincere and lying ways.

In the epistemic sense I think that justification roughly means that one uses other ideas as authorities (broad sense) as to why a particular idea is true. Like “I know that idea X is true because idea Y supports it”.
It is my understanding that one problem with justification (epistemic) is that justifying ideas leads to an infinite regress as each idea has to be justified by another idea.

I’m in the process of reading the links you and curi recommended to me to learn more.

Yeah – I’d say one’s justification for something is the collection of reasons for some thought, action, choice, whatever.

Re sincerity: Yes, a justification works for both cases. There’s no implication that a justification is true / honest / sincere / accurate / whatever.

(tangentially relevant if you haven’t read it: Lying · Elliot Temple)

I think “other ideas as authorities” is too narrow, but that mb is an artifact of your wording. Your example hits on something though – support.

I’ll try to make an example of justifying an idea (epistemically):

  • We know that light scatters in air (Rayleigh scattering) based on wavelength
  • We know that colors correspond to wavelength (shorter wavelengths scatter more)
  • So, roughly, whatever color the sky is must be the color with the shortest wavelength (out of red, green, blue)
  • The sky is red ATM (Note: see the below photo I took of a sunset recently; it’s not actually sunset when I’m writing this, but pretend it is)
  • Therefore red has the shortest wavelength

Everything I said up to the conclusion was correct. I justified the conclusion. But, ofc, we know the conclusion was not correct.

Maybe someone isn’t convinced – I could show more photographs (I have many, somewhere, of that particular sunset; and I could take more this evening, and tomorrow morning, and tomorrow evening, etc). I could justify the conclusion even more – indefinitely.

If you want, mb write a bit about what you think the problem is – with my method not the argument.

Note: JTB is like more than what I’ve described above, it’s a bit more subtle; I’m just talking about philosophical justification generally atm (which is broader than JTB and used in other schools of thought, too, like instrumentalism, and well most of them).

Ok. Thanks for the clarification on this.

I have read it in the past. I should read it again though.

I agree that it (authorities) is too narrow. My mistake. “Authority” is only one specific type of support. There are many other types of support as well which I left out by only picking one specific type.

Good idea. I will try to do that.

As I understand it, a problem with the above method (justification/justificationism) is that it is positive argument based. It tries to confirm/prove something. But we can’t infallibly confirm/prove things. The way that we learn is by eliminating errors. By justifying things we can’t eliminate errors. We need to use criticism to to eliminate errors, not justification.

A criticism of the above conclusion could be that the sky is not red but blue during the day. You could use the same method of justification during mid day and come to the conclusion that blue has the shortest wavelength out of red, green, and blue (which is factually correct).
These two conclusions contradict, so they can’t both be correct. We would need to try to rule one of these conclusions out (at least one - they could both be wrong). We can only do that by criticism in the form of critical argument or experiment when possible (such as in this case), not justification. Adding more positive arguments (justification) could never solve this contradiction.

Yes, though I have the sense that you’re parroting a bit (I don’t mean this as an insult, more that it’s mb an indication you have some depth to gain in your understanding).

Aside: I’m conscious this mb isn’t within the rules of #friendly@moderators are welcome to pull me up. TBH, I don’t know enough about the role that I am playing in this discussion to know if that was a good or bad thing to say.

Your answer could be more specific. Like there are things that – in following that method – I did in order to reach the conclusion (that I wanted to). Things like selecting evidence (both time of day and angle-from-horizon of photograph). A photo of a blue sky would be a crit, and so to would a photo of a night sky.

Note: I think I have more to say, will reply more tomorrow (getting late for me)

PS

Yes – if a method can show whatever you want it to (with little effort) then the method is worthless.

I agree. I am parroting a bit and have some depth to gain in my understanding. I did let it rest for ~a day and tried to reason about it from what I have recently read though.

It’s fine by me. You’re helping me understand and learn this better and I appreciate that. Hopefully it helps you learn better as well as you explain stuff and think of examples to use in this discussion.

Right. I think I gave an example of some selective evidence (the sentence starting with “A criticism of …”). But I did not identify it as a principle (“the principle of selective evidence” or something - just one single example of it). This might be an example of me lacking depth of understanding. Like I should have tried to think of principles and not just something specific to this very particular problem (sky thing).
Elliot talks about this kind of refute more problems of the same kind approach in video 4, “Rejecting the Burden of Proof”, of Yes or No Philosophy" (I listened to it during a walk some hours ago).

Ok.

By “this kind of refute more problems of the same kind approach” I mean general purpose criticism/ideas.

If you’re concerned you might not actually understand something, you should try discussing it applied to a concrete example from real life and see what people say about your example. If you can’t come up with such an example, that probably means you haven’t connected the abstractions you are dealing with to reality.

Don’t be too hard on yourself. It’s hard to think of principles without first having ideas that are specific to the situation (gotta start somewhere). Also, not everything has to be reduced to principles like that. (reasoning from first principles can be good, but reduction to first principles isn’t necessarily good)

Some quotes from About the Friendly category

In Friendly, posters choose what topics they want to discuss. Suppose someone posts a physics question. Don’t reply about how their question reveals that they’re dishonest, emotional or overreaching. Don’t criticize them for ending the discussion early or for not doing Paths Forward. Those would be topic changes away from their intended topic (physics). If you do want replies like that, put your topic in Unbounded.

Challenging, questioning and disagreeing with people is OK, but don’t subtly or indirectly undermine people. Don’t be passive-aggressive. Don’t hint at people being wrong or bad, while maintaining plausible deniability about whether or not you’re criticizing or attacking them. Don’t imply criticism without saying it openly or directly.

What follows is some quick thoughts from one of the admins - not the Authoritative Final Word by any measure

So, this current topic is named “Trying to understand justification better”. That’s @deroj’s goal. Suggesting that he could be parroting ideas about that topic points out a possible issue with his efforts to try to understand justification better. Therefore, it’s on topic. Also, the issue was raised in a direct and forthright way, and not in a subtle or indirect manner. Finally, @deroj says he appreciated it. So it’s no problem.

Dicta

One thing I will say is that there’s not an expectation that people get these issues 100% perfect. Making best efforts and being reasonable is the standard. You can also course correct based on feedback (both from the people you’re talking to and the admins).

In Friendly I think there should be a general norm of respecting the scope of the topics that people want to talk about.

OTOH, if someone posts in Unbounded and then is like “oh darnit I don’t actually want this unbounded crit”, well, it was misleading for them to post it where they did. It’s not an unfixable disaster (topics can be moved) but the error is on them in that case.

1 Like

Here’s an idea – instead of trying to think in terms of principles (which - generally for some topic - will get easier as you learn more, but will be harder at the start), why don’t you brainstorm all the things that might have been wrong about my method before? I’m sure there’s more than I’ve mentioned so far. If you can’t think of more than a few, that’s okay (still post those), you could try and pick another example of flawed reasoning (where someone is trying to justify something) and analyze that too. IMO you shouldn’t worry about going too deep atm, better to go wider and come up with as many problems as you can think of. Also, dw if you’re unsure whether something is a problem or not – write it down anyway b/c it’s a thing to discuss in that case. Plus, if you have like 8 criticisms of a method, and 4 of those criticisms don’t pan out, well you still have 4 legit crits!

One reason that I think this is good is b/c it avoids trying to see principles too early. That’s like prematurely optimizing your ideas. (If you think you found a principle, it’s good to write it down still, but don’t be afraid to let it go if it doesn’t work out.) The thing about principles is they must work generally, so it’s going to be an error prone process (most of the time) if you like come up with 2 things, come up with a principle, come up with a 3rd thing, fix the principle, come up with a 4th thing, fix the principle, … etc.

I don’t think all this abstraction discussion is necessary. It’d be better to just get started with the main event. E.g. read and comment on some of the relevant articles. That will either go well or it will run into problems, in which case discussion could be focused on stuff relevant to solving those problems.