(The parent is a long comment and makes several points, so I’m going to answer it in several parts. This is part 3.)
The central claims of the OP (as I understand it) are:
So, first all, it’s not clear to me that this is a good summary of the OP (in the sense that—it seems to me—it adds your interpretation to it, rather than representing the post directly). That being said, I’m not Benquo, so perhaps this is indeed what he meant. But regardless of any of that, let me go ahead and respond to these claims point by point:
There is a distinction between using tools and inventing them.
This is rather vague. A distinction? What is the nature of this distinction, exactly? I mean, this claim seems trivially true, in some sense, but I’m not sure how important it’s supposed to be, or how fundamental, etc. Elaborations (with details and examples) would help.
This distinction is reflected in explanations, as explanations vary in how much they improve the use of tools and how much they improve the invention of tools.
So, first of all, it seems to me like either there’s an implication that all human activity can be placed into one of these two categories (“using tools” vs. “inventing tools”). Or perhaps it’s only all human activity of some specific type? If so, what type would that be?
Or, otherwise, the question arises: do some explanations do things other than “improve the use of tools” or “improve the invention of tools”—either instead of, or in addition to? What other things might those be?
Also, is there any correlation between how much any given explanation improves the use of tools vs. how much it improves the invention of tools? Or is this a linear spectrum? Or are these totally orthogonal dimensions? And if there is a correlation, what is its causal origin? (And are these categories even sensible?)
And another question: how does the domain-specificity of an explanation interact with the degree to which “does this improve the use of tools” and “does this improve the invention of tools” even make sense as questions to ask?
The structure and content of an explanation is linked to how that explanation varies on those dimensions, and that there is a style of explanation that focuses on whys and connections that results in more improvement along the dimension of invention of tools.
This seems much too vague a claim for me to say anything more about it than what I’ve said above, re: #2.
Many explanations are solely judged on how well they improve tool use in a narrow dimension (because, perhaps, this is the only thing that can be verifiably tested) and one should expect this to lead to explanations that are deficient at improving tool invention.
There are actually several claims in here, which must be untangled before they can be addressed. My questions re: #2 seem like reasonable first steps toward untangling this question also.
It is desirable to have the ability to invent tools as well as just use them.
This is difficult to evaluate. Negate it, and we have:
“It is undesirable to have the ability to invent tools; we should only be able to use tools, not invent them.”
I have trouble imagining who would ever endorse this claim, so #5 is either an applause light… or, it’s an oblique way of suggesting that we should move further in some direction, on some purported spectrum.
Taken that way, we might interpret it as saying that we (for some value of “we”) currently have insufficient ability to invent tools, and should have more. In which case, it seems necessary to make, and defend, an explicit, positive claim about where on this purported spectrum we currently are, as well as a normative claim about where we ought to be. (The prerequisite for all of this, of course, would be establishing the structure of the purported spectrum in the first place, as I comment on above.)
That example [with the potatoes —SA] only engages with some of the points; it’s a demonstration of 1 and 5 more than it is 2, 3, or 4.
It is not a coincidence that #1 and #5 are, as I say above, the least interesting and most trivial of the claims (at least when taken at face value).
As an aside—though this is not (I think) terribly relevant to the OP—it does not seem to me like the “using tools vs. inventing tools” dichotomy (of which I am still rather skeptical) is all that natural a fit for characterizing your example with the potato mashing. (After all, you didn’t actually invent a new tool!) One could also describe it in some sort of “outside-the-box thinking” terms, or perhaps in terms of some sort of “analytical skills”[1], or perhaps in terms of “seeing the artificiality of purposes”[2], etc. We could have endless fun with the game of inventing plausible paradigms with which to describe this bit of clever thinking on your part… but I do not think the exercise would gain us any useful understanding.
[1] In a fairly literal sense of the word: we might say, perhaps, the you analyzed the act of using a potato masher to mash potatoes—that is, you decomposed the act into its constituent parts, such as “exerting pressure on the potatoes to crush them and deform their structure, breaking them up” and “using an instrument shaped so as to allow downward pressure to be exerted over a wide area”, etc.; and that this analysis, this skill of breaking-down, is what allowed you to synthesize your clever solution.
[2] Namely, of course, the fact that “the glass is for drinking out of” is not a property of the glass, and that the glass simply is the specific physical object that it is; it has no little XML tag attached, where its purpose is stored. We might, perhaps, claim that a keen sense of the artificiality of purpose is what allows one to perceive the possibility of unorthodox uses for designed artifacts.
So, first of all, it seems to me like either there’s an implication that all human activity can be placed into one of these two categories (“using tools” vs. “inventing tools”). Or perhaps it’s only all human activity of some specific type? If so, what type would that be?
I meant the latter, and the answer is going to be unsatisfying: the type is “using or inventing tools.” Specifically, note that “inventing” is a subcategory of “using,” worth separating out because it uses cognitively distinct labor (like the sort you refer to as ‘analyzing’). Then the questions are “okay, what fuels that cognitively distinct labor? What explanations make people better at analyzing?”.
This also seems related to all of the other questions you raise in this subsection, where it seems like you’re trying to expand the claim (consider the difference between “A and B are different” and “are you implying that everything is A or B?”, and consider the difference between “there are two dimensions” and “what are the statistical properties of the real world along those dimensions?”). I am sort of torn on this (as conversational technique), because it seems useful at exploring the thing, but also changes the direction of the conversation in a way that increases feelings of friction, or something.
In particular, I have only vague opinions on the statistical properties of explanations in the wild, and so the question puts me in something of a bind where either I share my vague opinions (which, if they get expanded, means I am even further aground, and if they get contradicted, the general point may be lost in the controversy) or somehow dismiss the question (which has its own share of drawbacks), and this bind is an example of the sort of friction this sort of thing can generate.
In which case, it seems necessary to make, and defend, an explicit, positive claim about where on this purported spectrum we currently are, as well as a normative claim about where we ought to be.
In the OP, the examples of this are Benquo’s mother and Benquo, in the context of yeast. In general, I don’t think I agree that such claims need to be explicit, and it’s not obvious to me that you have the right standards for ‘defend.’
I expect to be able to broadcast advice on how to lose weight and trust that readers have their own sense of their weight and their own sense of their desired weight and their own sense of their other tradeoffs, such that they judge my advice accordingly, rather than requiring that any advice on how to lose weight come packaged with disclaimers about anorexia and a discourse on what sort of measurements are actually connected to anything good. Those things should exist somewhere, of course—the culture should have ‘anorexia’ as a concept and discuss it sometimes, and should have concepts for things like bodyfat percentage—but requiring not just that they exist everywhere but also that anyone who touches on any part of the topic be an expert on the whole topic dramatically limits what can be said.
Connecting back: the discussion of leavening here reads to me as something like “don’t pretend that you know what leavening is in my presence, you ignoramus!”, which is a move that is occasionally sensible to make (if, say, Benquo claimed to know more about yeast than Said does, a battle of facts to settle the matter seems appropriate) but seems out of place here. [The reasons why it seems out of place here touch on controversies that are long to get into. I am not making a generic “this chills speech and that’s bad” argument as some speech should be chilled; I am instead making a claim of the form “this way to divide chilled and unchilled speech doesn’t line up with the broader goals of advancing the art of human rationality.”]
[This is just responding to points that are easy to respond to contained in the parent; my overall sense is “it might take a post to point at what’s going on here, and so I’m going to try to write that post instead of handle it here.”]
I don’t really follow most of what you say here about types of explanations, conversation directions, and feelings of friction, so I won’t respond to that part. Perhaps the post you mean to write will clarify things.
Concerning this bit specifically:
… requiring not just that they exist everywhere but also that anyone who touches on any part of the topic be an expert on the whole topic dramatically limits what can be said.
I think it’s often good to dramatically limit what can be said. I think that there are many cases where most of what can be said—in the absence of such limits—is nonsense; and we should try our outmost to ensure that nonsense cannot be said, and only not-nonsense can be said.
Connecting back: the function of the discussion of leavening here, reads to me as something like “don’t pretend that you know what leavening is in my presence, you ignoramus!“, which is a move that is occasionally sensible to make (if, say, Benquo claimed to know more about yeast than Said does, a battle of facts to settle the matter seems appropriate) but seems out of place here.
If, indeed, it is the case that Benquo doesn’t know what leavening is, then this is an excellent reason to distrust what he says in the OP. In that hypothetical case (as, absent a reply from Benquo, we still do not know whether it is the case in reality), the fact that it was in my presence that he pretended to that specific knowledge which he does not possess is fortuitous, as it allowed that lack of knowledge to be pointed out, and thus gave all readers the opportunity to reduce their credence in the OP’s claims. (One of the reasons why I so vehemently oppose insight-porn type posts is due to the Dan Brown phenomenon, especially combined with the “Gell-mann amnesia” effect: something can sound very much like real knowledge/expertise, but without an actual expert on hand to verify it, how do we know it isn’t just a pile of nonsense? We don’t, of course. See also Scott’s old post about epistemic learned helplessness, which talks about how explanations can be very convincing while being total nonsense.)
[The reasons why it seems out of place here touch on controversies that are long to get into. I am not making a generic “this chills speech and that’s bad” argument as some speech should be chilled; I am instead making a claim of the form “this way to divide chilled and unchilled speech doesn’t line up with the broader goals of rationality.“]
Well, as you say: some speech should be chilled. But I look forward to your more detailed commentary on this matter.
I appreciate the writing and clarity, but also (as we’ve gone over in a few past discussions) disagree on the object level. I ended up downvoting this comment because I think the position it defends is potentially quite damaging on a norms level, but wanted to make it clear that I do not disagree with the phrasing, method of communication, or fact that this comment was written.
(Of course, this highlights problems with karma serving multiple purposes that are sometimes at odds, which I am aware of and would still like to fix, but for now we have what we have)
(The parent is a long comment and makes several points, so I’m going to answer it in several parts. This is part 3.)
So, first all, it’s not clear to me that this is a good summary of the OP (in the sense that—it seems to me—it adds your interpretation to it, rather than representing the post directly). That being said, I’m not Benquo, so perhaps this is indeed what he meant. But regardless of any of that, let me go ahead and respond to these claims point by point:
This is rather vague. A distinction? What is the nature of this distinction, exactly? I mean, this claim seems trivially true, in some sense, but I’m not sure how important it’s supposed to be, or how fundamental, etc. Elaborations (with details and examples) would help.
So, first of all, it seems to me like either there’s an implication that all human activity can be placed into one of these two categories (“using tools” vs. “inventing tools”). Or perhaps it’s only all human activity of some specific type? If so, what type would that be?
Or, otherwise, the question arises: do some explanations do things other than “improve the use of tools” or “improve the invention of tools”—either instead of, or in addition to? What other things might those be?
Also, is there any correlation between how much any given explanation improves the use of tools vs. how much it improves the invention of tools? Or is this a linear spectrum? Or are these totally orthogonal dimensions? And if there is a correlation, what is its causal origin? (And are these categories even sensible?)
And another question: how does the domain-specificity of an explanation interact with the degree to which “does this improve the use of tools” and “does this improve the invention of tools” even make sense as questions to ask?
This seems much too vague a claim for me to say anything more about it than what I’ve said above, re: #2.
There are actually several claims in here, which must be untangled before they can be addressed. My questions re: #2 seem like reasonable first steps toward untangling this question also.
This is difficult to evaluate. Negate it, and we have:
“It is undesirable to have the ability to invent tools; we should only be able to use tools, not invent them.”
I have trouble imagining who would ever endorse this claim, so #5 is either an applause light… or, it’s an oblique way of suggesting that we should move further in some direction, on some purported spectrum.
Taken that way, we might interpret it as saying that we (for some value of “we”) currently have insufficient ability to invent tools, and should have more. In which case, it seems necessary to make, and defend, an explicit, positive claim about where on this purported spectrum we currently are, as well as a normative claim about where we ought to be. (The prerequisite for all of this, of course, would be establishing the structure of the purported spectrum in the first place, as I comment on above.)
It is not a coincidence that #1 and #5 are, as I say above, the least interesting and most trivial of the claims (at least when taken at face value).
As an aside—though this is not (I think) terribly relevant to the OP—it does not seem to me like the “using tools vs. inventing tools” dichotomy (of which I am still rather skeptical) is all that natural a fit for characterizing your example with the potato mashing. (After all, you didn’t actually invent a new tool!) One could also describe it in some sort of “outside-the-box thinking” terms, or perhaps in terms of some sort of “analytical skills”[1], or perhaps in terms of “seeing the artificiality of purposes”[2], etc. We could have endless fun with the game of inventing plausible paradigms with which to describe this bit of clever thinking on your part… but I do not think the exercise would gain us any useful understanding.
[1] In a fairly literal sense of the word: we might say, perhaps, the you analyzed the act of using a potato masher to mash potatoes—that is, you decomposed the act into its constituent parts, such as “exerting pressure on the potatoes to crush them and deform their structure, breaking them up” and “using an instrument shaped so as to allow downward pressure to be exerted over a wide area”, etc.; and that this analysis, this skill of breaking-down, is what allowed you to synthesize your clever solution.
[2] Namely, of course, the fact that “the glass is for drinking out of” is not a property of the glass, and that the glass simply is the specific physical object that it is; it has no little XML tag attached, where its purpose is stored. We might, perhaps, claim that a keen sense of the artificiality of purpose is what allows one to perceive the possibility of unorthodox uses for designed artifacts.
I meant the latter, and the answer is going to be unsatisfying: the type is “using or inventing tools.” Specifically, note that “inventing” is a subcategory of “using,” worth separating out because it uses cognitively distinct labor (like the sort you refer to as ‘analyzing’). Then the questions are “okay, what fuels that cognitively distinct labor? What explanations make people better at analyzing?”.
This also seems related to all of the other questions you raise in this subsection, where it seems like you’re trying to expand the claim (consider the difference between “A and B are different” and “are you implying that everything is A or B?”, and consider the difference between “there are two dimensions” and “what are the statistical properties of the real world along those dimensions?”). I am sort of torn on this (as conversational technique), because it seems useful at exploring the thing, but also changes the direction of the conversation in a way that increases feelings of friction, or something.
In particular, I have only vague opinions on the statistical properties of explanations in the wild, and so the question puts me in something of a bind where either I share my vague opinions (which, if they get expanded, means I am even further aground, and if they get contradicted, the general point may be lost in the controversy) or somehow dismiss the question (which has its own share of drawbacks), and this bind is an example of the sort of friction this sort of thing can generate.
In the OP, the examples of this are Benquo’s mother and Benquo, in the context of yeast. In general, I don’t think I agree that such claims need to be explicit, and it’s not obvious to me that you have the right standards for ‘defend.’
I expect to be able to broadcast advice on how to lose weight and trust that readers have their own sense of their weight and their own sense of their desired weight and their own sense of their other tradeoffs, such that they judge my advice accordingly, rather than requiring that any advice on how to lose weight come packaged with disclaimers about anorexia and a discourse on what sort of measurements are actually connected to anything good. Those things should exist somewhere, of course—the culture should have ‘anorexia’ as a concept and discuss it sometimes, and should have concepts for things like bodyfat percentage—but requiring not just that they exist everywhere but also that anyone who touches on any part of the topic be an expert on the whole topic dramatically limits what can be said.
Connecting back: the discussion of leavening here reads to me as something like “don’t pretend that you know what leavening is in my presence, you ignoramus!”, which is a move that is occasionally sensible to make (if, say, Benquo claimed to know more about yeast than Said does, a battle of facts to settle the matter seems appropriate) but seems out of place here. [The reasons why it seems out of place here touch on controversies that are long to get into. I am not making a generic “this chills speech and that’s bad” argument as some speech should be chilled; I am instead making a claim of the form “this way to divide chilled and unchilled speech doesn’t line up with the broader goals of advancing the art of human rationality.”]
[This is just responding to points that are easy to respond to contained in the parent; my overall sense is “it might take a post to point at what’s going on here, and so I’m going to try to write that post instead of handle it here.”]
I don’t really follow most of what you say here about types of explanations, conversation directions, and feelings of friction, so I won’t respond to that part. Perhaps the post you mean to write will clarify things.
Concerning this bit specifically:
I think it’s often good to dramatically limit what can be said. I think that there are many cases where most of what can be said—in the absence of such limits—is nonsense; and we should try our outmost to ensure that nonsense cannot be said, and only not-nonsense can be said.
If, indeed, it is the case that Benquo doesn’t know what leavening is, then this is an excellent reason to distrust what he says in the OP. In that hypothetical case (as, absent a reply from Benquo, we still do not know whether it is the case in reality), the fact that it was in my presence that he pretended to that specific knowledge which he does not possess is fortuitous, as it allowed that lack of knowledge to be pointed out, and thus gave all readers the opportunity to reduce their credence in the OP’s claims. (One of the reasons why I so vehemently oppose insight-porn type posts is due to the Dan Brown phenomenon, especially combined with the “Gell-mann amnesia” effect: something can sound very much like real knowledge/expertise, but without an actual expert on hand to verify it, how do we know it isn’t just a pile of nonsense? We don’t, of course. See also Scott’s old post about epistemic learned helplessness, which talks about how explanations can be very convincing while being total nonsense.)
Well, as you say: some speech should be chilled. But I look forward to your more detailed commentary on this matter.
I appreciate the writing and clarity, but also (as we’ve gone over in a few past discussions) disagree on the object level. I ended up downvoting this comment because I think the position it defends is potentially quite damaging on a norms level, but wanted to make it clear that I do not disagree with the phrasing, method of communication, or fact that this comment was written.
(Of course, this highlights problems with karma serving multiple purposes that are sometimes at odds, which I am aware of and would still like to fix, but for now we have what we have)