Just as commenters shouldn’t have assumed Eliezer’s factual observation was an argument in favor of regulation,
But did they assume it? Or did they conclude it based on inferences from Eliezer’s comment and the broader context?
To recast that in more local-jargon, Bayesian terms… how high was their prior probability that Eliezer was making an argument in favor of regulation, and how much evidence in favor of that proposition was the comment itself, and did they over-weight that evidence?
Beats me, I wasn’t there. I might not be able to tell, even if I had been there. But saying they “assumed” it in this context connotes that their priors were inappropriately high.
I’m not sure that connotation is justfiied, either in the specific case you quote Eliezer as discussing, or in the general case you and he treat it as illustrative of.
Maybe, instead, they were overweighting the evidence provided by the comment itself.
Or maybe they were weighting the evidence properly and arriving at, say, a .7 confidence that Eliezer was making an argument in favor of regulation, and (quite properly) made their bet as though that was the case… and turned out, in this particular case, to be wrong, as they should expect in 3 out of 10 cases.
you shouldn’t assume the suspected Blue’s observation is a pro-moon shot or anti-Green argument.
Sure, agreed. But here again, not assuming it doesn’t preclude me from concluding it.
When I choose to make an utterance, I am not only providing you with the utterance’s propositional content. I am also providing you with the information entailed by the fact that I chose to utter it.
When you make inferences about my motives from that information, you might of course be mistaken. But that doesn’t mean you shouldn’t make such inferences.
Agreed. But as I said, it’s not clear to me that inferring the propositions under discussion is invalid or incorrect, so to the extent that “invalidly or incorrectly infer” is what’s meant, I’m skeptical of the claim. Ditto for “jump to a conclusion” for the most common connotations of that phrase.
When I wrote the comment it seemed more charitable to give the claim the reading under which I agree with it, and then point out the more complicated reality of which it is a narrow slice, than to give the claim the reading under which I simply doubt that it’s true. In retrospect, though, I’m not sure it was.
Either way, though, my main point is that inferring that someone is making a covert argument while seeking to maintain the social cover of just making a factual observation is not necessarily unjustified in cases like these.
You weren’t there. You can’t reconstruct what it was like to be there. But you can read his comment. It contains the word “tradeoff” four times. Can you suggest what disclaimers he should have used instead?
(but the comments responding to Eliezer seem pretty reasonable to me.)
Can you suggest what disclaimers he should have used instead?
Let’s assume for the sake of comity that I can’t. What follows?
To address your broader question, though: it seems likely to me that there is no wording which reliably causes observers to believe that I’m genuinely just making a factual observation and that I’m not covertly implying any arguments, since I can’t think of any way of preventing people who are covertly implying arguments from using the same wording, which will shortly thereafter cause clever observers to stop trusting that wording.
This certainly includes bald assertions like “Hey, guys, I’m genuinely just making a factual observation here and totally NOT covertly implying any arguments, OK?” which even unsophisticated deceivers know enough to use, but it also covers more sophisticated variations.
That said, it also seems likely to me that for any given audience there exists wording that will manipulate that audience into believing I’m genuinely just making a factual observation, and a sufficiently skilled manipulator can find that wording. I don’t claim to be such a manipulator. (Of course, if I were, it would probably be in my best interests not to claim to be.)
Then again, such a manipulator could presumably do this even when that belief is false.
The approach I usually endorse in such cases is to not worry about it and concentrate on more generally behaving in a trustworthy way, counting on observant members of the community to recognize that and to consequently trust me to not be playing rhetorical games. (That’s not to say I always succeed, nor that I never play rhetorical games.) In other words, I count on the cultivation of personal reputation over iterated trials.
Of course, deceivers of all stripes similarly count on the cultivation of personal reputation over iterated trials.
Expensive signaling helps here, of course, but isn’t always an option.
The more important question is whether people should state hostile inferences based on usually flimsy evidence. I think vocally pointing out intentions behind factual claims is a very effective way to discourage rational discussion and cause mindkill because the rate of false positives is so high. Manufacturing plausible deniability by just stating facts works precisely because deniability in such a case should be plausible to have any relevant discussion at all.
To take your comment as an example… on one level, it’s a series of claims. “X is the more important question.” “Y is an effective way to discourage rational discussion.” “The rate of false positives in Y is very high.” Etc. And I could respond to it on that level, discussing whether those claims are accurate or not. And that seems to be the kind of discussion you’re encouraging.
Had you instead responded by saying “The average rainfall in Missouri is 3.5 inches per year” I could similarly discuss whether that claim is accurate or not.
But that would be an utterly bizarre response. Why would it be bizarre? Because I would have no idea what the intention behind citing that fact could possibly be. Your comment, by contrast, seems to have a fairly clear intention behind it, so it’s not bizarre at all.
So far, I don’t think I’ve said anything in the least bit controversial. (If you disagree with any of the above, probably best to pause here and resolve that disagreement before continuing.)
Continuing… so, OK. You have certain intentions in making the comment you made… call those intentions I1. I have inferred certain intentions on your part… call those I2. And, as above, were I to lack a plausible I2, I would be utterly bewildered by the whole conversation, as in the Missouri rainfall example… which I’m not.
Now… if I understand your view correctly, you believe that if I articulate I2 I will effectively discourage rational discussion and cause mindkill, because I’m likely to be mistaken… that is, I2 is not likely to equal I1. It’s better, on your view, for me to continue holding I2 without articulating it.
Yes? Or have I misunderstood your view?
If I’ve understood your view correctly, I disagree with it completely.
I tried to focus on people attacking negative intentions/connotations. I was expressing myself poorly and my comment had a lot of hidden assumptions. My comment was not even wrong. Your response is clear and helpful, thanks. I’m not sure I can improve upon my original comment, but here are some thoughts on the matter:
I think it would be useful to categorize intentions/connotations further. I see no problem in articulating hostile intentions behind a comment rudely stating that someone is fat for example. I think the reason for this is that the connotations of that kind of a statement are common knowledge and high probability. If you disapprovingly point out such connotations, nobody can claim that you’re trying to sneak them into the other person’s comment to dismiss it unfairly.
Then again I think there’s this category of statements where it seems to me that connotations can vary wildly. Even if you have a good reason to think that some particular connotation is the most probable, it’s just one option among many. Here the rate of false positives will be high. I feel in such situations attacking one connotation over another seems like a dishonest way to dismiss a statement.
I acknowledge that situational factors complicate matters further.
Even if you have a good reason to think that some particular connotation is the most probable, it’s just one option among many. Here the rate of false positives will be high.
Sure, that’s true. We might disagree about how high my confidence in a particular most-probable-interpretation of the motives behind a particular statement can legitimately be, but it’s clear that for some statements that confidence will be fairly low.
I feel in such situations attacking one connotation over another seems like a dishonest way to dismiss a statement.
Do you have any sense of why you feel this way?
For example, do you believe it is a dishonest way to dismiss a statement? Or just that it seems that way? (Seems that way to whom?)
But did they assume it?
Or did they conclude it based on inferences from Eliezer’s comment and the broader context?
To recast that in more local-jargon, Bayesian terms… how high was their prior probability that Eliezer was making an argument in favor of regulation, and how much evidence in favor of that proposition was the comment itself, and did they over-weight that evidence?
Beats me, I wasn’t there.
I might not be able to tell, even if I had been there.
But saying they “assumed” it in this context connotes that their priors were inappropriately high.
I’m not sure that connotation is justfiied, either in the specific case you quote Eliezer as discussing, or in the general case you and he treat it as illustrative of.
Maybe, instead, they were overweighting the evidence provided by the comment itself.
Or maybe they were weighting the evidence properly and arriving at, say, a .7 confidence that Eliezer was making an argument in favor of regulation, and (quite properly) made their bet as though that was the case… and turned out, in this particular case, to be wrong, as they should expect in 3 out of 10 cases.
Sure, agreed. But here again, not assuming it doesn’t preclude me from concluding it.
When I choose to make an utterance, I am not only providing you with the utterance’s propositional content. I am also providing you with the information entailed by the fact that I chose to utter it.
When you make inferences about my motives from that information, you might of course be mistaken. But that doesn’t mean you shouldn’t make such inferences.
The same goes for your hypothetical Blue.
People often say “assume” when they mean “jump to a conclusion” or “invalidly or incorrectly infer”. That seems to be what’s meant here.
Agreed. But as I said, it’s not clear to me that inferring the propositions under discussion is invalid or incorrect, so to the extent that “invalidly or incorrectly infer” is what’s meant, I’m skeptical of the claim. Ditto for “jump to a conclusion” for the most common connotations of that phrase.
When I wrote the comment it seemed more charitable to give the claim the reading under which I agree with it, and then point out the more complicated reality of which it is a narrow slice, than to give the claim the reading under which I simply doubt that it’s true. In retrospect, though, I’m not sure it was.
Either way, though, my main point is that inferring that someone is making a covert argument while seeking to maintain the social cover of just making a factual observation is not necessarily unjustified in cases like these.
You weren’t there. You can’t reconstruct what it was like to be there. But you can read his comment. It contains the word “tradeoff” four times. Can you suggest what disclaimers he should have used instead?
(but the comments responding to Eliezer seem pretty reasonable to me.)
Let’s assume for the sake of comity that I can’t.
What follows?
To address your broader question, though: it seems likely to me that there is no wording which reliably causes observers to believe that I’m genuinely just making a factual observation and that I’m not covertly implying any arguments, since I can’t think of any way of preventing people who are covertly implying arguments from using the same wording, which will shortly thereafter cause clever observers to stop trusting that wording.
This certainly includes bald assertions like “Hey, guys, I’m genuinely just making a factual observation here and totally NOT covertly implying any arguments, OK?” which even unsophisticated deceivers know enough to use, but it also covers more sophisticated variations.
That said, it also seems likely to me that for any given audience there exists wording that will manipulate that audience into believing I’m genuinely just making a factual observation, and a sufficiently skilled manipulator can find that wording. I don’t claim to be such a manipulator. (Of course, if I were, it would probably be in my best interests not to claim to be.)
Then again, such a manipulator could presumably do this even when that belief is false.
The approach I usually endorse in such cases is to not worry about it and concentrate on more generally behaving in a trustworthy way, counting on observant members of the community to recognize that and to consequently trust me to not be playing rhetorical games. (That’s not to say I always succeed, nor that I never play rhetorical games.) In other words, I count on the cultivation of personal reputation over iterated trials.
Of course, deceivers of all stripes similarly count on the cultivation of personal reputation over iterated trials.
Expensive signaling helps here, of course, but isn’t always an option.
The more important question is whether people should state hostile inferences based on usually flimsy evidence. I think vocally pointing out intentions behind factual claims is a very effective way to discourage rational discussion and cause mindkill because the rate of false positives is so high. Manufacturing plausible deniability by just stating facts works precisely because deniability in such a case should be plausible to have any relevant discussion at all.
I don’t think I agree.
To take your comment as an example… on one level, it’s a series of claims. “X is the more important question.” “Y is an effective way to discourage rational discussion.” “The rate of false positives in Y is very high.” Etc. And I could respond to it on that level, discussing whether those claims are accurate or not. And that seems to be the kind of discussion you’re encouraging.
Had you instead responded by saying “The average rainfall in Missouri is 3.5 inches per year” I could similarly discuss whether that claim is accurate or not.
But that would be an utterly bizarre response. Why would it be bizarre? Because I would have no idea what the intention behind citing that fact could possibly be. Your comment, by contrast, seems to have a fairly clear intention behind it, so it’s not bizarre at all.
So far, I don’t think I’ve said anything in the least bit controversial. (If you disagree with any of the above, probably best to pause here and resolve that disagreement before continuing.)
Continuing… so, OK. You have certain intentions in making the comment you made… call those intentions I1. I have inferred certain intentions on your part… call those I2. And, as above, were I to lack a plausible I2, I would be utterly bewildered by the whole conversation, as in the Missouri rainfall example… which I’m not.
Now… if I understand your view correctly, you believe that if I articulate I2 I will effectively discourage rational discussion and cause mindkill, because I’m likely to be mistaken… that is, I2 is not likely to equal I1. It’s better, on your view, for me to continue holding I2 without articulating it.
Yes? Or have I misunderstood your view?
If I’ve understood your view correctly, I disagree with it completely.
I tried to focus on people attacking negative intentions/connotations. I was expressing myself poorly and my comment had a lot of hidden assumptions. My comment was not even wrong. Your response is clear and helpful, thanks. I’m not sure I can improve upon my original comment, but here are some thoughts on the matter:
I think it would be useful to categorize intentions/connotations further. I see no problem in articulating hostile intentions behind a comment rudely stating that someone is fat for example. I think the reason for this is that the connotations of that kind of a statement are common knowledge and high probability. If you disapprovingly point out such connotations, nobody can claim that you’re trying to sneak them into the other person’s comment to dismiss it unfairly.
Then again I think there’s this category of statements where it seems to me that connotations can vary wildly. Even if you have a good reason to think that some particular connotation is the most probable, it’s just one option among many. Here the rate of false positives will be high. I feel in such situations attacking one connotation over another seems like a dishonest way to dismiss a statement.
I acknowledge that situational factors complicate matters further.
Sure, that’s true. We might disagree about how high my confidence in a particular most-probable-interpretation of the motives behind a particular statement can legitimately be, but it’s clear that for some statements that confidence will be fairly low.
Do you have any sense of why you feel this way?
For example, do you believe it is a dishonest way to dismiss a statement? Or just that it seems that way? (Seems that way to whom?)