Eliezer does have a post in which he talks about doing what you call conceptual analysis more-or-less as you describe and why it’s worthwhile. Unfortunately, since that’s just one somewhat obscure post whereas he talks about tabooing words in many of his posts, when LWrongers encounter conceptual analysis, their cached thought is to say “taboo your words” and dismiss the whole analysis as useless.
The ‘taboo X’ reply does seem overused. It is something that is sometimes best to just ignore when you don’t think it aids in conveying the point you were making.
When I try that, I tend to get down-votes and replies complaining that I’m not responding to their arguments.
I don’t know the specific details of the instances in question. One thing I am sure about, however, is that people can’t downvote comments that you don’t make. Sometimes a thread is just a lost cause. Once things get polarized it often makes no difference at all what you say. Which is not to say I am always wise enough to steer clear of arguments. Merely that I am wise enough to notice when I do make that mistake. ;)
In the example he does start with a word, namely ‘art’, then uses our intuition to get a set of examples. This is more-or-less how conceptual analysis works.
I disagree. Suppose after proposing a definition of art based to the listed examples, someone produced another example that clearly satisfied our intuitions of what constituted art but didn’t satisfy the definitions. Would Eliezer:
a) say “sorry despite our intuitions that example isn’t art by definition”, or
b) conclude that the example was art and there was a problem with the definition?
He’s not trying to define art in accord with on our collective intuitions, he’s trying to find the simplest boundary around a list of examples based on an individual’s intuitions.
I would argue that the list of examples in the article is abbreviated for simplicity. If there is no single clear simple boundary between the two sets, one can always ask for more examples. But one asks an individual and not all of humanity.
He’s not trying to define art in accord with on our collective intuitions, he’s trying to find the simplest boundary around a list of examples based on an individual’s intuitions.
I would argue he’s trying to find the simplest coherent extrapolation of our intuitions.
Why do we even care about what specifically Eliezer Yudkowsky was trying to do in that post? Isn’t “is it more helpful to try to find the simplest boundary around a list or the simplest coherent explanation of intuitions?” a much better question?
Focus on what matters, work on actually solving problems instead of trying to just win arguments.
The answer to your question is “it depends on the situation”. There are some situations in which are intuitions contain some useful, hidden information which we can extract with this method. There are some situation in which our intuitions differ and it makes sense to consider a bunch of separate lists.
But, regardless, it is simply the case that when Eliezer says
“Perhaps you come to me with a long list of the things that you call “art” and “not art”″
and
“It feels intuitive to me to draw this boundary, but I don’t know why—can you find me an intension that matches this extension? Can you give me a simple description of this boundary?”
he is not talking about “our intuitions”, but a single list provided by a single person.
(It is also the case that I would rather talk about that than whatever useless thing I would instead be doing with my time.)
Eliezer’s point in that post was that there are more and less natural ways to “carve reality at the joints.” That however much we might say that a definition is just a matter of preference, there are useful definitions and less useful ones. The conceptual analysis lukeprog is talking about does call for the rationalist taboo, in my opinion, but simply arguing about which definition is more useful as Eliezer does (if we limit conceptual analysis to that) does not.
Eliezer does have a post in which he talks about doing what you call conceptual analysis more-or-less as you describe and why it’s worthwhile. Unfortunately, since that’s just one somewhat obscure post whereas he talks about tabooing words in many of his posts, when LWrongers encounter conceptual analysis, their cached thought is to say “taboo your words” and dismiss the whole analysis as useless.
The ‘taboo X’ reply does seem overused. It is something that is sometimes best to just ignore when you don’t think it aids in conveying the point you were making.
When I try that, I tend to get down-votes and replies complaining that I’m not responding to their arguments.
I don’t know the specific details of the instances in question. One thing I am sure about, however, is that people can’t downvote comments that you don’t make. Sometimes a thread is just a lost cause. Once things get polarized it often makes no difference at all what you say. Which is not to say I am always wise enough to steer clear of arguments. Merely that I am wise enough to notice when I do make that mistake. ;)
I do not think that he is describing conceptual analysis. Starting with a word vs. starting with a set of objects makes all the difference.
In the example he does start with a word, namely ‘art’, then uses our intuition to get a set of examples. This is more-or-less how conceptual analysis works.
But he’s not analyzing “art”, he’s analyzing the set of examples, and that is all the difference.
I disagree. Suppose after proposing a definition of art based to the listed examples, someone produced another example that clearly satisfied our intuitions of what constituted art but didn’t satisfy the definitions. Would Eliezer:
a) say “sorry despite our intuitions that example isn’t art by definition”, or
b) conclude that the example was art and there was a problem with the definition?
I’m guessing (b).
He’s not trying to define art in accord with on our collective intuitions, he’s trying to find the simplest boundary around a list of examples based on an individual’s intuitions.
I would argue that the list of examples in the article is abbreviated for simplicity. If there is no single clear simple boundary between the two sets, one can always ask for more examples. But one asks an individual and not all of humanity.
I would argue he’s trying to find the simplest coherent extrapolation of our intuitions.
Why do we even care about what specifically Eliezer Yudkowsky was trying to do in that post? Isn’t “is it more helpful to try to find the simplest boundary around a list or the simplest coherent explanation of intuitions?” a much better question?
Focus on what matters, work on actually solving problems instead of trying to just win arguments.
The answer to your question is “it depends on the situation”. There are some situations in which are intuitions contain some useful, hidden information which we can extract with this method. There are some situation in which our intuitions differ and it makes sense to consider a bunch of separate lists.
But, regardless, it is simply the case that when Eliezer says
“Perhaps you come to me with a long list of the things that you call “art” and “not art”″
and
“It feels intuitive to me to draw this boundary, but I don’t know why—can you find me an intension that matches this extension? Can you give me a simple description of this boundary?”
he is not talking about “our intuitions”, but a single list provided by a single person.
(It is also the case that I would rather talk about that than whatever useless thing I would instead be doing with my time.)
Eliezer’s point in that post was that there are more and less natural ways to “carve reality at the joints.” That however much we might say that a definition is just a matter of preference, there are useful definitions and less useful ones. The conceptual analysis lukeprog is talking about does call for the rationalist taboo, in my opinion, but simply arguing about which definition is more useful as Eliezer does (if we limit conceptual analysis to that) does not.