I strongly disagree with your “not allowed to care” claim at the end, and I note with some interest that the link you offer in support of it is something you yourself wrote, and whose comment section consists mostly of other rationalists disagreeing with you. (I also note that it got a lot of upvotes, and one possible explanation for that is that a lot of people agreed with you who didn’t say so in the comments.)
We get to define our words however the hell we like. There are many ways in which one set of definitions can be better or worse than another, one of which is that ceteris paribus definitions are better when they reduce average message length. But it’s not the only one, and it shouldn’t be the only one. (If that were the only criterion, then we should completely redesign our language to eliminate all redundancy. There are many reasons why that would be bad; for instance, errors would have worse consequences, and the language would likely be harder to learn and use because it would match our brain-hardware worse. Note that neither of those reasons is a matter of reducing message length.)
In any case, you have not by any means proved that your proposed change would reduce average message length. I think the nearest you get is where you say
But I claim that broad “algorithmic” senses of “mental” words like intent often are more relevant and useful for making sense of the world than the original, narrower definitions that were invented by humans in the context of dealing with other humans, because the universe in fact does not revolve around humans.
and I think it’s very clear that falls far short of demonstrating anything. (For several reasons, of which I will here mention only one: although indeed the universe does not revolve around humans, it happens that most of the time when we are discussing things like “intent” we are talking about humans, and definitions / linguistic conventions that lead to less clarity when talking about humans and more clarity when talking about other entities that might have something like “intent” are not obviously better.)
(Regretfully, I’m too busy at the moment to engage with most of this, but one thing leapt out at me—)
I note with some interest that the link you offer in support of it is something you yourself wrote, and whose comment section consists mostly of other rationalists disagreeing with you
(Why write a lengthy and potentially controversial piece if you know you haven’t time to engage with responses? But --)
[EDITED to add:] Of course, maybe you have plenty of time to engage with other responses and something about mine specifically or about me specifically makes you value such engagement particularly little. In that case there’d be no particular inconsistency. I don’t know of any reason why my comments, specifically, would be not worth engaging with—but then I wouldn’t, would I?
The statement “as an aspiring epistemic rationalist I am not allowed to do X” can be interpreted in three ways. (1) “Not doing X is part of what the word ‘rationalist’ means.” (2) “Among rationalists, X is morally prohibited.” (3) “X is in some fashion objectively wrong for everyone, and it happens that rationalists pay particular attention to that sort of wrongness.”
The behaviour of others who consider themselves rationalists is relevant to #1 because the meaning of a word is determined by how it is actually used. It is relevant to #2 because what is prohibited in a given community is determined by the opinions of that community as a whole. It is only tangentially relevant to #3, and I suspect that #3 is your actual meaning; but (a) prima facie #1 and #2 are also possible, and (b) even with #3 I think one function of “as an aspiring epistemic rationalist” in what you wrote is to encourage readers who also think of themselves that way to feel bad about disagreeing, which I think they shouldn’t and are less likely to if aware that either your usage of “rationalist” or your opinion about what “rationalists” are allowed to do is highly idiosyncratic.
Why write a lengthy and potentially controversial piece if you know you haven’t time to engage with responses?
Look, I have a dayjob and I’m way behind on chores! I had a blog post idea that I’d been struggling with on-and-off for months, that I finally managed to finish a passable draft of and email to prereaders Sunday night. My prereaders thought it sucked (excerpts: “should ideally be a lot tighter. 2⁄3 the word count?”, “Don’t know why [this post] exists. It seems like you have something to say, so say it”), but on Monday night I was able to edit it into something that I wasn’t embarrassed to shove out the door—even if it wasn’t my best work. I have a dayjob videoconference meeting starting in two minutes. Can I maybe get back to you later???
Sure, whatever works for you. (Including not getting back to me later, of course. If what I wrote came across as trying to impose an obligation then I put it badly.) I hope the videoconference went well.
So, I’m still behind on chores (some laundry on the floor, some more in the dryer, that pile of boxes in the living room, &c.), but allow me to quickly clarify one thing before I get to sleep. (I might have a separate reply for the great-grandparent later.)
the meaning of a word is determined by how it is actually used
Right, and the same word can have different meanings depending on context if it gets used differently in different contexts. Specifically, I perceive the word rationalist as commonly being used with two different meanings:
(1) anyone who seeks methods of systematically correct reasoning, general ways of thinking that produce maps that reflect the territory and use those maps to formulate plans that achieve goals
(2) a member of the particular social grouping of people who read Less Wrong. (And are also likely to read Slate Star Codex, be worried about existential risk from artificial intelligence, be polyamorous and live in Berkeley, CA, &c.)
I intended my “as an aspiring epistemic rationalist” to refer to the meaning rationalist(1), whereas I read your “other rationalists disagreeing with you” to refer to rationalist(2).
An analogy: someone who says, “As a Christian, I cannot condone homosexual ‘marriage’” is unlikely to be moved by the reply, “But lots of Christians at my liberal church disagree with you”. The first person is trying to be a Christian(1) as they understand it—one who accepts Jesus Christ as their lord and savior and adheres to the teachings of the Bible. The consensus of people who are merely Christian(2)—those who belong to a church—is irrelevant if that church is corrupt and has departed from God’s will.
Hope this helps.
one function of “as an aspiring epistemic rationalist” in what you wrote is to encourage readers who also think of themselves that way to feel bad about disagreeing
I mean, they should feel bad if and only if feeling bad helps them be less wrong.
Yup, all understood. (I think in practice any given use of the word is likely to have a bit of both meanings in it, with or without concomitant equivocation.)
[EDITED to add:] Maybe it’s worth saying a little about your analogy. To whatever extent the analogy does more than merely illustrate your distinction between rationalist-1 and rationalist-2 (and I take it it is intended to do a little more, since the distinction was perfectly clear without it), it seems that you see yourself as being in something like the position of the Hypothetical Evangelical, asking us all “do ye not therefore err, because ye know not the Sequences, neither the power of rationality?”. That of course is why I dedicated most of my comment to arguing against your actual claim, that those who seek truth through clear thinking must choose their concept-boundaries without any considerations other than minimizing average description length. And my position is a bit like one I remember holding not infrequently in the past, when I was (alas) a moderately-evangelical Christian: I can see how you see the Sequences as supporting your position, but I don’t think that’s the only or the best way to interpret the relevant bits of the Sequences, and to whatever extent Eliezer was saying the same thing as you I’m afraid I think that Eliezer was wrong. (Yes, I am suggesting, tongue somewhat but not wholly in cheek, a parallel between some Christians’ equivocation between “God’s will” and “what is written in the bible” and your appeal to the authority of Eliezer’s posts about word and concepts when arguing for what seem to me inadvisably-extreme positions on how rationalists should use words.)
Incidentally, I’m aware that that’s now twice in a row that I’ve responded very briefly and then edited in more substantive comments. I promise I’m not doing it out of any wish to deceive or anything like that. It’s just that sometimes I’m right on the fence about how much it’s worth saying.
I mean, I agree that pretending Levels 2+ don’t exist may not be a good strategy for getting Level 1 discourse insofar as it’s hard to coordinate on … but maybe not as hard as you think?
How sure are you about that? Other things Zack’s written seem to me to argue for a rather similar position without any obvious sign of irony. Zack, if you’re reading this, would you care to comment?
I strongly disagree with your “not allowed to care” claim at the end, and I note with some interest that the link you offer in support of it is something you yourself wrote, and whose comment section consists mostly of other rationalists disagreeing with you. (I also note that it got a lot of upvotes, and one possible explanation for that is that a lot of people agreed with you who didn’t say so in the comments.)
We get to define our words however the hell we like. There are many ways in which one set of definitions can be better or worse than another, one of which is that ceteris paribus definitions are better when they reduce average message length. But it’s not the only one, and it shouldn’t be the only one. (If that were the only criterion, then we should completely redesign our language to eliminate all redundancy. There are many reasons why that would be bad; for instance, errors would have worse consequences, and the language would likely be harder to learn and use because it would match our brain-hardware worse. Note that neither of those reasons is a matter of reducing message length.)
In any case, you have not by any means proved that your proposed change would reduce average message length. I think the nearest you get is where you say
and I think it’s very clear that falls far short of demonstrating anything. (For several reasons, of which I will here mention only one: although indeed the universe does not revolve around humans, it happens that most of the time when we are discussing things like “intent” we are talking about humans, and definitions / linguistic conventions that lead to less clarity when talking about humans and more clarity when talking about other entities that might have something like “intent” are not obviously better.)
(Regretfully, I’m too busy at the moment to engage with most of this, but one thing leapt out at me—)
I don’t think that should be interesting. I don’t care what so-called “rationalists” think; I care about what’s true.
(Why write a lengthy and potentially controversial piece if you know you haven’t time to engage with responses? But --)
[EDITED to add:] Of course, maybe you have plenty of time to engage with other responses and something about mine specifically or about me specifically makes you value such engagement particularly little. In that case there’d be no particular inconsistency. I don’t know of any reason why my comments, specifically, would be not worth engaging with—but then I wouldn’t, would I?
The statement “as an aspiring epistemic rationalist I am not allowed to do X” can be interpreted in three ways. (1) “Not doing X is part of what the word ‘rationalist’ means.” (2) “Among rationalists, X is morally prohibited.” (3) “X is in some fashion objectively wrong for everyone, and it happens that rationalists pay particular attention to that sort of wrongness.”
The behaviour of others who consider themselves rationalists is relevant to #1 because the meaning of a word is determined by how it is actually used. It is relevant to #2 because what is prohibited in a given community is determined by the opinions of that community as a whole. It is only tangentially relevant to #3, and I suspect that #3 is your actual meaning; but (a) prima facie #1 and #2 are also possible, and (b) even with #3 I think one function of “as an aspiring epistemic rationalist” in what you wrote is to encourage readers who also think of themselves that way to feel bad about disagreeing, which I think they shouldn’t and are less likely to if aware that either your usage of “rationalist” or your opinion about what “rationalists” are allowed to do is highly idiosyncratic.
Look, I have a dayjob and I’m way behind on chores! I had a blog post idea that I’d been struggling with on-and-off for months, that I finally managed to finish a passable draft of and email to prereaders Sunday night. My prereaders thought it sucked (excerpts: “should ideally be a lot tighter. 2⁄3 the word count?”, “Don’t know why [this post] exists. It seems like you have something to say, so say it”), but on Monday night I was able to edit it into something that I wasn’t embarrassed to shove out the door—even if it wasn’t my best work. I have a dayjob videoconference meeting starting in two minutes. Can I maybe get back to you later???
Sure, whatever works for you. (Including not getting back to me later, of course. If what I wrote came across as trying to impose an obligation then I put it badly.) I hope the videoconference went well.
So, I’m still behind on chores (some laundry on the floor, some more in the dryer, that pile of boxes in the living room, &c.), but allow me to quickly clarify one thing before I get to sleep. (I might have a separate reply for the great-grandparent later.)
Right, and the same word can have different meanings depending on context if it gets used differently in different contexts. Specifically, I perceive the word rationalist as commonly being used with two different meanings:
(1) anyone who seeks methods of systematically correct reasoning, general ways of thinking that produce maps that reflect the territory and use those maps to formulate plans that achieve goals
(2) a member of the particular social grouping of people who read Less Wrong. (And are also likely to read Slate Star Codex, be worried about existential risk from artificial intelligence, be polyamorous and live in Berkeley, CA, &c.)
I intended my “as an aspiring epistemic rationalist” to refer to the meaning rationalist(1), whereas I read your “other rationalists disagreeing with you” to refer to rationalist(2).
An analogy: someone who says, “As a Christian, I cannot condone homosexual ‘marriage’” is unlikely to be moved by the reply, “But lots of Christians at my liberal church disagree with you”. The first person is trying to be a Christian(1) as they understand it—one who accepts Jesus Christ as their lord and savior and adheres to the teachings of the Bible. The consensus of people who are merely Christian(2)—those who belong to a church—is irrelevant if that church is corrupt and has departed from God’s will.
Hope this helps.
I mean, they should feel bad if and only if feeling bad helps them be less wrong.
Yup, all understood. (I think in practice any given use of the word is likely to have a bit of both meanings in it, with or without concomitant equivocation.)
[EDITED to add:] Maybe it’s worth saying a little about your analogy. To whatever extent the analogy does more than merely illustrate your distinction between rationalist-1 and rationalist-2 (and I take it it is intended to do a little more, since the distinction was perfectly clear without it), it seems that you see yourself as being in something like the position of the Hypothetical Evangelical, asking us all “do ye not therefore err, because ye know not the Sequences, neither the power of rationality?”. That of course is why I dedicated most of my comment to arguing against your actual claim, that those who seek truth through clear thinking must choose their concept-boundaries without any considerations other than minimizing average description length. And my position is a bit like one I remember holding not infrequently in the past, when I was (alas) a moderately-evangelical Christian: I can see how you see the Sequences as supporting your position, but I don’t think that’s the only or the best way to interpret the relevant bits of the Sequences, and to whatever extent Eliezer was saying the same thing as you I’m afraid I think that Eliezer was wrong. (Yes, I am suggesting, tongue somewhat but not wholly in cheek, a parallel between some Christians’ equivocation between “God’s will” and “what is written in the bible” and your appeal to the authority of Eliezer’s posts about word and concepts when arguing for what seem to me inadvisably-extreme positions on how rationalists should use words.)
Incidentally, I’m aware that that’s now twice in a row that I’ve responded very briefly and then edited in more substantive comments. I promise I’m not doing it out of any wish to deceive or anything like that. It’s just that sometimes I’m right on the fence about how much it’s worth saying.
The claim at the end is a sarcastic poke at a particular straw rationality also hinted at by Simulacra levels etc.
I mean, I agree that pretending Levels 2+ don’t exist may not be a good strategy for getting Level 1 discourse insofar as it’s hard to coordinate on … but maybe not as hard as you think?
How sure are you about that? Other things Zack’s written seem to me to argue for a rather similar position without any obvious sign of irony. Zack, if you’re reading this, would you care to comment?