curi Quotes

ON THE PURPOSE OF LIFE:

"in short the purpose of life is to do some productive things that you like."

ON THE 'INERTNESS' OF EVIDENCE:

"another point that’s subtle is that evidence is *inert*. it doesn’t actually favor theories or hint at anything or even contradict theories. evidence is only used in arguments, never directly, b/c evidence has to be interpreted to have a meaning. so even in science, you can’t just do bayesian updating, first you need arguments, and to deal with arguments. then you can do something kinda like bayesian updating, in limited contexts, governed by arguments dealt with according to CR."

ON GENUINE EXPLANATIONS BEING STATEMENTS OF HOW SOMETHING IS CAUSED:

"Actually without a causal *explanation* you shouldn't conclude anything. (you need some concept of how the causation could work, that makes sense. you can never ever go purely by correlation)."

"Are you familiar with Russel’s chicken? The sun rises every morning until it doesn’t. We have a causal understanding of why it rises and what would cause it to stop, not just repeated observations."

"here’s an example of the importance of *why/how*: A funny story is the one with the London taxi drivers. A survey has pointed out a positive and significant correlation between the number of accidents and wearing coats. It was assumed that coats could hinder movements of drivers and be the cause of accident. A new law was prepared to prohibit drivers to wear coats when driving. Finally another study pointed out that people wear coats when it rains! Rain was the hidden factor common to wearing coat and accident frequency. so it’s crucial to know why coats correlate to accidents in order to decide what to do about it.

ON THE BENEFITS OF MODIFYING IDEAS TO BE MORE UNIVERSAL AND CONNECTIVE:

"It is good to fit theories into more general frameworks, or alter them to be more universal, or connect them with other ideas we have. Doing this is interesting in its own right; it's part of how we learn. But it also helps us correct errors. When our ideas don't mesh together nicely, that is a sign they could be improved. And the more generally they apply, the more varied examples we can try them out on, which will help reveal hidden flaws."

ON THE LIBERAL CONCEPTION OF FREEDOM:

"the situation of living with one's rights protected, and without any limits on your actions other than not violating the rights of others, is called freedom. in short, it's freedom from violence, and freedom to do non-violent actions."

ON SOLIPSISM:

"If you're a solipsist, the world behaves AS IF the standard view is true. So you still have to understand the standard view, then you throw in an extra complication."

BETTING ON THEORIES WHEN THE EVIDENCE IS 'FUZZY':

gjm: "Imagine that you're counting neutrino detections. There are vast numbers of neutrinos but the experimental apparatus fails to detect most of them; so far as we know, it does so at random. Each theory (A and B) predicts some rate of neutrino production and hence some rate of neutrino detection. After, say, a year of observation, theory A says we should expect the number of neutrinos detected to be Poisson-distributed with mean 10 and theory B says they should be Poisson-detected with mean 20. We actually see 12 neutrinos. These results are, say, 90% likely if A is true, and 10% likely if B were true. Your colleague offers to bet at 30:1 odds that A is correct. Will you take his bet?"

Temple: "the answer is basically to tentatively accept a meta-theory which mentions probability (in the sense of quantified uncertainty with certain mathematical properties), and reject all other theories. this doesn’t introduce probability into epistemology directly – all the ideas are judged in a binary way. both “I know A” and “I know B” are *wrong* things to think here, and the right thing to think is “I don’t know if A or B in the usual sense of e.g. refuting all the rivals of A, but, given a great deal of context, and incomplete information, the best guess at this point is A, and the odds required to bet on A over B are...” (for simplicity, the bet specifies that if it turns out to be anything other than A or B, then no one pays out). this allows the use of e.g. bayesian math, just not inside epistemology."

ON THE INAPPROPRIATE APPLICATION OF ODDS:

"keep in mind it's probabilities of ideas being true that we object to, not all ideas which mention probabilities. e.g.:

1) "in Holdem poker with 4 cards of a suit after the flop you have a 34.97% chance to make a flush on the turn or river"

2) the probability that claim (1) is true is 99.7%.

claim (1) is fine and claim (2) is crap."

"Odds apply to issues like physical events. Odds are a reasonable way to think about the possibility of dying in a plane crash [...] You could die in a plane crash, or not. It could go either way, so odds make some sense. But either current cryo methods (assume perfusion etc go well) preserve the necessary information, or they don't. That can't go either way, there's a fact of reality one way or the other."

ON UNIVERSAL THEORIES:

"universal propositions, this says, attribute some property to some class (universal means it applies to ALL of the class). example:

class: uranium-235

property: half-life is 704 million years

why is this universal? b/c it's talking about how ALL uranium-235 behaves.

half-life is probabilistic. so what?"

ON THEORIES THAT MAKE EXCEPTIONS:

"whether theories make exceptions is important. exceptions can get in the way of criticism. one way to deal with criticism is to add an exception for each criticism. theories with no exceptions, or only exceptions that have a good explanation, are best and are more vulnerable to criticism (including involving counter examples)."

AN EXAMPLE OF A UNIVERSAL THEORY THAT MENTIONS PROBABILITY:

""Rolling 5 6-sided dice, rather than 4, increases the risk of getting a sum above 10."

This is a universal theory. it applies to all dice, not just red ones. it applies to all people, not just hindu dice-throwers. it applies at all locations, not just on Earth."

ON WHAT COERCION IS:

"coercion is when people use force against you and take away your choices."

ON THE VALUE OF PRINCIPLES AND WHY EXCEPTIONS ARE BAD:

"Arbitrary exceptions to principles are also bad because they partially defeat the purpose of principles.

The point of the principle is to help you in multiple situations. it's an idea you can re-use in multiple cases. this is good because it saves time/effort. it's good b/c it helps you think or act consistently across situations (makes life more predictable and manageable).

the whole point of a principle is that some idea is good and important to use across a range of situations/questions/problems.

exceptions mean not using it and using something else.

what else? i don't know. often some ad hoc crap. but regardless: if something else should be used, there is a flaw in the principle which ought to be addressed. if an exception really is needed, fix the principle or reject it!

put another way: exceptions to principles can mean not using a good idea where it applies. they can also mean continuing to use a bad idea most of the time and trying to cover up how it doesn't work by ignoring it in a few key cases.

exceptions can be used to lower the impact of your good ideas ... which sucks. and exceptions can be used to lower the impact of your bad ideas ... which can help them stick around and slip through the cracks more."

ON THE CAUSES OF CRIME:

"The problem is people who want to hurt each other. A few of the causes relating to that problem in the USA include:

1) the war on drugs (bad idea)

2) racial issues (especially with blacks)

3) bad public schools, especially in certain neighborhoods and inner cities

4) left wing ideas which place less respect and value on individual human lives"

"It also sounds like you're saying people commit crimes because they're poor. But lots of poor people don't commit crimes. So I think you should look more closely at what the difference is between the different poor people who are or aren't criminals."

ON GUNS:

"I think the way it SHOULD work is if someone attacks you, and you're doing a reasonable job defending yourself, and someone [a bystander] gets shot, the ATTACKER is guilty of causing that shooting."

ON WHEN TO DECIDE WHEN SOMETHING EXISTS:

"we should deem things to exist if, and only if, they are needed in our best explanations."

ON THE ABILITY OF THINGS TO BE EXPLANATIONS OF THEMSELVES:

"you're mistaken about "nothing can be explained only in terms of itself". that is a vague and bad argument. two counter examples:

1) the totality of everything (the multiverse, every idea, the rules of logic, EVERYTHING including all abstract things) contains its own explanation(s)

2) "When quined provides an example of a quine. A quine is a statement preceded by a quotation of itself, see the book _Godel, Escher, Bach_ for more info." When quined provides an example of a quine. A quine is a statement preceded by a quotation of itself, see the book _Godel, Escher, Bach_ for more info.

here i embedded an explanation of a quine in a quine. so it explains itself. (you could, if you wanted, put a longer more thorough explanation of a quine in the same spot, including a full dictionary to explain English, the full text of _GEB_, and whatever else you think is required.)"

ON SUPPOSEDLY DIFFERENT CONCEPTIONS OF TRUTH:

Erin Minter asked:

> one conception of truth says that truth is objective and it exists in reality outside of our minds. There is a truth to what actually exists in reality and we can’t control the truth just by our thinking or by wishing it be different. In order to consider an idea of ours as “true”, it needs to match this objective reality.

> another conception of truth says truth has to do with successfully resolving contextual idea-conflicts in our minds. Truth is for problem solving (which is an internal mental process). This does not require a true idea to actually work in reality.

> Do these 2 conceptions conflict? If we are able to create a TRUE idea to resolve a mental idea-conflict which does not work in reality (does not match reality), then I think they do conflict.

Temple responds:

"these aren't different conceptions of truth. they are different categories of questions one can ask. you can ask questions about the world itself. and you can also ask questions about things like: "given Bob's life situation and limited knowledge, what action should he decide to do?" and then if Bob thinks a store closes at 10pm, but really it's 9pm, he might go there after it closes (oops) but that could still have been the right decision to make given the information he had."

ON HOW TRUE SOLUTIONS DON'T HAVE TO WORK IN REALITY:

"Suppose: I have a choice to make. Given the available information, the best choice for me is X. I do X. X doesn't work. In retrospect, Y would have worked, which I'd also considered. X was true – it was the contextually correct answer to the question of what I should choose at that time. X did not work."

ON CORROBORATION:

""Corroborate" is defined as something like "try to falsify, but fail"."

ON THE PURPOSE OF LAWS:

"I think the proper purpose of laws is to basically to outlaw force in order to maintain a peaceful society. I don't think laws should be designed to try to be helpful or pleasant or guide anyone. Any government action comes with big downsides, so laws should be a bare minimum of stopping the really bad stuff (force). (This requires a nuanced understanding of what force is. Alternatively for a quick approximation, just include also fraud and threat of force.)"

A REPLY RE: OCCAM'S RAZOR:

">If a theory does not reduce the number of elements needed to understand the world, it is not serving is function.

Basically, I agree. But that is a critical, negative argument! Not induction or positive support."

ON THE FALLIBILITY OF DEDUCTION:

"conclusions NEVER necessarily follow, they can only fallibly follow according to some non-refuted guesses. a lot of thinking surrounding "deduction" is infallibilist error."

ON WHAT BELIEF IS:

"Correctly understood, it is having an idea."

ON HOW *ARGUMENTS* AFFECT IDEAS:

"Rather than supporting or damaging ideas directly, evidence is used in arguments which refer to the evidence. It's the argument (which refers to evidence), not the evidence (itself), which affects the status of ideas. Put another way, evidence is meaningless without ideas that interpret it, and the role of those ideas is under appreciated."

"To get from moss on that side of the tree--the observational evidence--to knowing which way is north, you need a complex argument invoking lots of concepts. The evidence not only has to be interpreted from percepts to concepts like moss, it also has to then be used in an argument which refers to the moss to make a claim about where north is.

So you argue like: North is that way because i see moss on that side, and moss likes the shade, and we're in the northern hemisphere, and it'll get more shade on the north, and there's nothing around to mess that up and no signs of human intervention, and it's the same on many trees.

In that argument, the observational evidence is used, but it's nothing like the majority of the content. There's all this other stuff too. Evidence properly conceived is always used like that, although yes it often makes a crucial difference."

"In my layman's understanding of courtrooms, lawyers introduce a bunch of stuff called evidence, then separately make arguments about what to make of the evidence. Often opposing lawyers will offer rival arguments about the same evidence. That's basically how I see it: the evidence doesn't speak for itself, you have to make arguments that refer to it."

ON 'DEGREES':

"stuff has degrees when an explanation defines or refers to some degrees.

raw, pure epistemology doesn't have any degrees. they aren't there at the fundamental beginnings of the field. they come in later.

like you are trying to choose a restaurant for dinner. the basic thing is you need a single solution with no criticism, which resolves any conflicts or problems about which to pick that you know of.

how do you pick a restaurant? there are many, many ways. ONE way is you might DEFINE some ways of measuring restaurant quality in degrees, then look at the results of estimating those quantities. so you CREATE YOUR OWN degrees, in the context of having an explanation of why those degrees matter (e.g. you want to go to the restaurant that scores highest on a particular measure you designed to meet your preferences well)."

ON THE NONEXISTENCE OF 'BETTER' OR 'WORSE' SOLUTIONS:

"Concepts of solving a problem better or worse are mixing up issues:

- the problem itself defines one or more measures, which different solutions score differently on (score better and worse)

- the actual human choices involved, if approached rationally, mean deciding some particular approach does solve the problem (YES rather than NO – boolean issue) and doing that. (the fact that some other options would solve the same problem, and are in some sense "worse", does not prevent each individual option from either solving or not solving the problem, booleanly)"

ON WHY THERE CAN BE MULTIPLE TRUE SOLUTIONS TO HUMAN PROBLEMS:

"human problems have multiple solutions because humans are not INFINITELY demanding and picky".

ON ARBITRARILY DENYING REACH:

"It's a mistake to deny reach arbitrarily. Ideas always have a right amount of reach. Ideas about electrons can/should apply to all electrons in the whole multiverse. Or ideas that apply to replicators apply to all replicators, inherently, but not to towels. Explanations about conjectures and refutations apply to all conjectures and all refutations. This is because the logic of the explanations apply to those cases. E.g. the explanations make use of aspects of conjectures (e.g. they are a type of idea), but not aspects of potatoes. So they apply to conjectures but not potatoes. They have an amount of reach that makes sense. Or consider axis tilting and seasons. Does that apply on other planets orbiting other suns? Yes. If they have axis tilting then we're going to see some of the same heat distribution effects. Saying it wouldn't work, because that's a different solar system, would be an unexplained denial of the explanations of how axis tilting and seasons work. The explanations never make use of which solar system they are in, so they don't just apply to one solar system."

ON THE POSSIBLE CLASSIFICATIONS OF IDEAS:

"Passing a test doesn't make something true. The real categories are:

- refuted by criticism

- not refuted by criticism

Testing is one type of criticism. 'Untestable' is not some special thing. All ideas can be criticized. Differentiating testable ideas is important in some limited contexts, but not in philosophy in general. Passing one test is not a special occasion that transforms an idea from undecided to true. That's completely wrong. What we have to do is consider ideas tentatively undecided if we have some pending criticism we haven't resolved yet. Such as we have an idea for a test, or an argument, that we think it's important to address, but we haven't done it yet. A more important milestone is when people run out of criticisms (including tests) they want to attempt. This may not last very long, but surviving the initial criticisms is a good start. Once an idea gets that far, for the first time it can say it's dealt with all the criticism anyone has thought of, rather than there are outstanding criticism attempts pending resolution."

ON WEIGHTING:

"they also want to ignore that they have no criticism of something, but say its merits are outweighed anyway. (but if it's actually missing some important merit, why not just criticize that inadequacy?)"

"Tentatively accept ideas when you have exactly one non-refuted candidate. Stop trying to assign ideas scores/weights of any kind. Then all the problems go away. Whatever factor was inspiring an increase or decrease in weighting either can or can’t inspire a criticism. If it can, make the criticism and go from there. If it can’t, it was worthless. And if something is refuted by a criticism, saying it still has some weight/status/justification instead of none is just a way to irrationally try to ignore criticism."

ON DESCRIPTIONS OF PROBLEMS:

"describing a problem is solving a problem (a problem of description, of understanding another problem and how to describe it)."

ON WANTING TO SOLVE PROBLEMS:

"It’s also important to note, in the context of problem solving, that there is a presumption in the concept of a “problem” that we want to solve it. If I don’t want to change something then that means I don’t regard it as a problem."

ON CLASSIFYING AN IDEA AS 'TRUE':

"Further, holding ideas to be true, or at least truthlike, isn't a bad thing. Fallibilists can and should do that. We have knowledge about what is true or not, and we should say so, even if we may be mistaken and have to change our positions later."

"Or if you don't like "truthlike" you could write "good", "high status", "highly evaluated", "the tentative winner in an adjudication of ideas", or whatever else. Whichever of those is the end goal, you could try to achieve it with either method: justification or criticism. The problem with justification is that it's a bad method of achieving any of those goals, not that it is inherently coupled with the wrong goal."

ON GETTING UNSTUCK WHEN WE NEED TO ACT:

"What you can do is ask questions in arbitration like, "Given we think we won't solve problems X, Y and Z within our resource constraints, what should we do?" That question can be answered without solving problems X, Y or Z, and its answer can be a successful win/win arbitration outcome.

As with everything, it's open to criticism, e.g. a side might think X actually can be solved within the resource constraints. Then all sides might be able to agree, for example, to try to solve X, but also to set up a backup plan in case that doesn't work.

If an arbitration seems particularly hard relative to the resources available, a longer exclusion list can be proposed. By setting things aside as necessary, arbitration can succeed in the short term."

AN EXAMPLE OF HOW ONE MIGHT GET UNSTUCK IN ORDER TO ACT:

"If theory X is false, the following theory may still be true: "X is false but also a good approximation in the following circumstances..." Using that second theory would not be using X, and it would not be using a false theory. Creating a second theory like this is one way we can sometimes solve the problem. The second theory does of course need to be exposed to criticism itself, and it may get refuted, but it may not."

ON DEUTSCH'S HARD-TO-VARY CRITERION FOR THE GOODNESS OF EXPLANATIONS:

"The right approach to epistemology is focused more on criticism. Judge ideas not by some favored criterion but by open-ended imaginative criticism. I think you can't swap the sun and moon because I have a criticism of that swap. The way hard to vary gets its plausibility, I think, is the unstated constraint is not to vary ideas in any way you have criticism of. It sneaks criticism into the equation. This is good in that it works better than a less critical approach, but bad as a theory of epistemology because it hides away the most important thing -- criticism -- instead of explaining and emphasizing the central role of criticism."

ON HUMAN VS. NON-HUMAN PROBLEMS:

"Wanting something is a human problem! Interests are a human issue!

A non-human problem is, *with no human context*: 2+2=?

But 2+2=? on *your math test* is a human problem (yours). Or if you want to know the answer, that is a human problem."

"Problem-1: (we might call it "human problem"): "a matter or situation regarded as unwelcome or harmful and needing to be dealt with and overcome"

Problem-2: (we might call it an "abstract problem"): "a thing that is difficult to achieve or accomplish"

ON WHAT A PROBLEM-SITUATION (=CONTEXT) IS:

"The problem situation is the context. it includes the context the problem came up in, and what is known about the problem and about solving it. every choice is tied to a problem situation by definition. choices come up in situations. whatever that situation the choice came up in is a 'problem situation'."

"The context, or 'problem situation', is not just a set of problems. It also has other stuff, like a list of your relevant skills and resources you could use."

"'problem situation' is, roughly, the Popperian words for talking about 'context' (which is Objectivism's term). a problem is, like, a part of the context you're looking to do something about."

"> so context is pretty much like all the relevant, active aspects which need to be considered (including not only any relevant ideas and preferences in your mind, but other relevant aspects of reality such as the laws of physics).

background knowledge and resources, such as books sitting quietly on a shelf that you could reference, are not 'active aspects' of your situation. but they are part of the context. and 'context' includes irrelevant stuff. that would be irrelevant context. it's still your context, your situation. normally you try not to mention what you judge isn't relevant, but that doesn't stop it from being part of your situation. your judgements about what's relevant, and what isn't, is another part of the context."

"> and there would be a *set* of problems included in this context.

yes though the context may be "i have a bunch of stuff in my life in general. books, meals, friends, etc. and problems too. and solutions – solved problems. but right now, i'm focusing on this one problem i want to solve. (which i want to solve without messing up any of my important solutions without a replacement.)""

ON WHAT A CRITICISM IS:

"A criticism both identifies some aspect of an idea and explains why it's bad. The explanation is necessary because one has to say *why* the thing is bad. The identification is necessary to point out what's being criticized."

"a criticism is an explanation about WHY an idea doesn't work well."

ON THE CONTEXTUALITY OF KNOWLEDGE:

"whether something is knowledge is contextual. knowledge is contextual. in one context, today, something is knowledge. then you refute it and in that new context it isn’t knowledge, it isn’t viable for solving any problems. but it still was – and always and forever was and will be – knowledge in that older context where it did have problem-solving power."

#14903 Everything after the blue part is curi's writing. I just messed up on quoting each of those lines in green like I should have. The blue quote is from the Wikipedia page on mirror neurons I believe.

Four years ago, in December 2015, curi wrote a short essay titled Criticism. It gives reasons why people don't like criticism and reasons why they do.

The essay has lots of quotable sections. Here are just a few of my favorites:

> Criticism sucks when it **makes** you do things. When it **controls** your life. When it's in charge, bossing you around.

> Criticism goes badly when people feel they have **no choice**. People need to make their own choices and use criticism to help, *not* get pressured and lose control over their decision making.

> People will also externalize the criticism to the **messenger** who told it to them. So they feel like that *other person* is trying to control them, make them do things, boss them around, etc.

> *People like criticism when they are in control of their life.* Then it's useful information they can use to solve problems they're already working on.

> When people have *proper interests* – the unlimited, unbounded, freely pursued beginning of infinity style of interest – then criticism is a gift.

> Nothing is wrong with happiness. The problem is aiming for it. It should be an indirect consequence of a good life. Trying to get happiness directly is either meaningless (because we only do what we think is moral), or it means we sometimes do things that we do not think are moral.

- <https://curi.us/1168>