Yeah, as a physician who also has chronic pain, this just about encapsulates the important refutation of the concept. Pain is a fraught concept, and the idea that a philosopher doing a Gedankenexperiment is going to have anything at all useful to say about it is not reasonable. Pain is much more about emotional salience which is very much not a primitive. Not to say infants don't experience a physical analogue of what adults refer to as pain, but by the time someone is expressing abstract thoughts about "pain" they've moved well beyond this.
This gets to something that is really a problem today. These "realisms" are a modern form of vitalism and they are, to a degree, pernicious. Pain is usually caused by physical precedents, and this talk of qualia and inherent nonphysical properties can mislead people into thinking that they should suffer with fixable pain. That's not to endorse the "all pain is a medical emergency" fallacy that led to the opioid crisis in any way, but there are a subset of people inclined by psychology to stoicism who get exposed to this kind of modernized vitalism and think that they shouldn't try to mitigate things that are giving them pain in ways that alter their lives and limit their goals.
Love your writing, and I love your incessant questioning of every assumption. Is there some root to all this philosophy writing relying on seeming?
When I think of bloggers in the rationality sphere talking about real world topics, there is often a lot of uncertainty and many times they'll run with an assumption ("X SEEMS to be the case, so we'll run with it for this analysis even though there's little/no data" ).
However, similar-sounding seeming (BB does it a lot, for example) is here applied as part of an armchair argument looking for some philosophical truth, say the stance-independence of moral facts. Are there just a lot of sloppy philosophers trying to make it as a writer after graduation? ChatGPT doesn't seem to think their thought has much basis in prior analytic philosophy canon, but I've read basically zero and couldn't say.
This reminds me of two things one is short and pithy. Hobbes said their were two sciences were we make everything and so admit of demonstrative knowledge (sure inference from first principles):
Geometry therefore is demonstrable for the lines and figures from which we reason are drawn and described by ourselves and civil philosophy is demonstrable because we make the commonwealth ourselves. (EW VII.184)
(This quote is apparently from Six Lessons to the Professors of Mathematiques.)
This overestimates the regularity or surety of demonstration we can achieve in civil philosophy.
More substantively it reminds me of the distinction made famous by Einstein between two types of scientific theories. Constructive theories as in statistical mechanics were a theory is developed from a simple hypothesis about some underlying mechanism. So this seems roughly similar to your top down concept. To this Einstein contrasts theories of principle like Einstein's own theory relativity or thermodynamics, which simply mathematically codify empirically observed generalizations. So this would roughly correspond to your bottom up concepts.
Einstein elaborates this distinction to elucidate issues with relativity in a popular account from 1919, "Time, Space, and Gravitation"
The basic distinction was already made by the 19th century Scottish physicist William John Macquorn Rankine in 1855 and had been discussed by various people including Pierre Duhem, which is where I assume Einstein got it from. Rankine contrasts hypothetical method of theorizing (that posit some unsee mechanism) against abstractive method (that identifies sets of things that have been observed by the senses to have common properties). Instead of hypothetical Morris Cohen and Ernest Nagel say physical and instead of abstractive they say mathematical.
I learned of Rankine from happening to read a little of Cohen and Nagel's 1936 book (they don't actually connect Rankine and Einstein, but its clear the exposition must be related, I guess Duhem is the likely connection because Einstein famously read some Duhem, Einstein could easily have read Rankine himself I spose):
I have expanded this reference to Einstein and others thoughts about scientific theories into a post. This is only tangentially related to the initial post, but this seems like a good place to link to it anyway. https://substack.com/home/post/p-173908686
Excellent piece! I agree with pretty much all of it. I share your skepticism of the opaque/transparent concept distinction and find the claims about being "super justified" about the nature of your own internal states extremely dubious, to say the least. Though for slightly different reasons. Our access to our internal states is just as theory-laden, fallible and subject to revision as any putative belief about the external world, Cartesian dogmatism notwithstanding. I'm inclined to think there isn't actually such thing as concepts that 'reveal essences' at all, and whether you count as knowing what a given concept refers to, depends on the (largely empirical) background knowledge you have. Though, the points you make here are also interesting and make me want to read into ordinary language philosophy which I've held off for a while!
This is a precise piece and I suspect (though I hold this tentatively) that the top-down/bottom-up distinction you’ve drawn may be doing even more than you’ve said here.
It may actually be the case that there are two distinct populations of language users who differ in what functions as a token versus a type in their processing. For one population **world-token processors** tokens are real-world phenomenological referents and words are then transparent pointers to felt experience. For this population ‘Types’ are fuzzy categories, bell curves of grouped, similar phenomenology.
For a second population **word-token processors** tokens are the words themselves. Precise, discrete, sharp-edged lexemes. Types for us are patterns of use, distributional roles, operators rather than felt categories. The phenomenological undergirding that justifies word usage for the first population is. I suspect, largely unavailable as a first-person justification for us.
And yes I suspect I’m in the second population and perhaps you may be also. And I further suspect we may be in the minority.
What makes this potentially interesting rather than merely taxonomic is a layering observation. There may be something like three consecutive layers: phenomenological experience, words that point at that experience, and the operator structure of how those words behave grammatically. World-token processors have transparent access to the first two layers — the felt experience that then generates the word that points to it. Word-token processors are, I suspect, largely out of phase with the first layer and find themselves instead needing a process, maybe even a hack to know when to use the words. This process or operator later is the third and we use it precisely because the first layer isn’t generating the meaning.
Being out of phase in this way may put a small minority in the position of seeing the operator structure clearly while finding the phenomenological grounding that most people take as obvious somewhat opaque. The parsimonious account isn’t that one population is reasoning correctly and the other isn’t. It’s that each population has transparent access to different layers of the same structure.
If this is right, your top-down/bottom-up distinction may be tracking something real not just about concepts but about the processors using them. I’d be curious whether any of this resonates.
"I don’t think there are concepts that (for some mysterious reason) are a bit voyeuristic and strip down so we can see all their tidbits, while others are more coy and waltz about clothed, revealing only an ankle here or a sultry gaze over there."
I believe that would make the concepts exhibitionists. We're the voyeurs, hoping to catch a glimpse of some top-down/bottom-up action.
Yeah this discussion reminds me of other things that you and I have spoken about before. The only thing I disagree with is when you say that "analytic philosophy goes astray" because of its failures to recognize these sorts of things. I would say that analytic philosophy EXISTS primarily because of these failures, and whether they're mistakes or not, they are very important parts of the basement-level presuppositions that make analytic philosophy what it is. In other words, if you don't share these presuppositions about language and thought, there is surprisingly little for you to do in mainstream analytic philosophy, because most of it depends on the kind of things you were talking about.
I think in the past you and I have called it a "residual Platonism" about concepts. As far as I can tell, mainstream analytic philosophy is thoroughly Platonist about concepts, in the sense that it takes it for granted that concepts have some special "essence" that we can "discover," and that grasping the content of a concept is a true and significantly meaningful cognitive achievement, kind of like popping your head outside Plato's cave and seeing by the light of the sun. It's like a genuine DISCOVERY to find out something about a particular concept, instead of, as you said, a simple recognition of what we say or stipulate. Inasmuch as analytic philosophy has articles of faith, this bit about concepts is one of them.
I remember having a discussion about this in one of my graduate seminars. We were reading an ethics paper with various thought experiments that involved the concept of a "killing," and various intuitions pushed people to say that one thing counted as a "killing" and another didn't. And it's like the professor felt that there was truly something meaningful to be discovered, the essence of the concept, if we could only sort through all the different factors. But in reality, a "killing" is whatever we want it to be, and so if we wanted a ham sandwich to count as a killing, it would count as a killing. There's nothing further to discover about concepts than that a lot of the time.
I think this is a little overstating the case against analytic philosophy. Analyzing even purely stipulative concepts can result in genuine discoveries in just the same way that mathematics can. It's often not clear at all what a collection of formal stipulations implies.
There is, however, a big problem with the use of thought experiments to discover the content of a concept.
I guess if by "discovery" you mean something like "pragmatic insight into how we could construct a concept differently in order to better serve our ends," then I would agree. I can't see how we could ever get more than that by analysis.
I'm thinking of how carrying out deductions based on formal definitions can prove statements we didn't know those formal definitions committed us to. If we stipulated different axiomatizations of set theory, we wouldn't thereby immediately know what downstream theorems were entailed. We'd have to reason through it.
Great write up. "Transparent concepts" are seriously problematic. I especially like the note about slipping between the ostensive definition and a stipulated definition. The formalized concept possesses its features essentially because it's a formalization, and it *does not* thereby follow that the ostensive concept actually has those features.
If I can't doubt that pain is bad, it doesn't follow that *this* pain is certainly bad--it follows that *if this is indeed pain*, then it's bad. But perhaps what I take to be pain isn't pain, just as you say.
This strikes me as a variant of the fixed points argument for moral realism. Seeing as that argument is extremely weak, this one doesn’t do much better. I’ve never been a big fan of these stipulative moves that realists make, and they certainly have little motivational force.
In the original post, the author claims
“When I stub my toe, I immediately recognise the badness of my pain.” I interpret this to mean something like “I have perception, in the same way as hearing, seeing, etc that pain is (normatively) bad.” I also find moral perception arguments to be pretty weak, as they venture into the realm of empirics, where I’ve seen no evidence to support that people literally and directly perceive moral properties.
Yeah, as a physician who also has chronic pain, this just about encapsulates the important refutation of the concept. Pain is a fraught concept, and the idea that a philosopher doing a Gedankenexperiment is going to have anything at all useful to say about it is not reasonable. Pain is much more about emotional salience which is very much not a primitive. Not to say infants don't experience a physical analogue of what adults refer to as pain, but by the time someone is expressing abstract thoughts about "pain" they've moved well beyond this.
This gets to something that is really a problem today. These "realisms" are a modern form of vitalism and they are, to a degree, pernicious. Pain is usually caused by physical precedents, and this talk of qualia and inherent nonphysical properties can mislead people into thinking that they should suffer with fixable pain. That's not to endorse the "all pain is a medical emergency" fallacy that led to the opioid crisis in any way, but there are a subset of people inclined by psychology to stoicism who get exposed to this kind of modernized vitalism and think that they shouldn't try to mitigate things that are giving them pain in ways that alter their lives and limit their goals.
Total banger! 🔥
Love your writing, and I love your incessant questioning of every assumption. Is there some root to all this philosophy writing relying on seeming?
When I think of bloggers in the rationality sphere talking about real world topics, there is often a lot of uncertainty and many times they'll run with an assumption ("X SEEMS to be the case, so we'll run with it for this analysis even though there's little/no data" ).
However, similar-sounding seeming (BB does it a lot, for example) is here applied as part of an armchair argument looking for some philosophical truth, say the stance-independence of moral facts. Are there just a lot of sloppy philosophers trying to make it as a writer after graduation? ChatGPT doesn't seem to think their thought has much basis in prior analytic philosophy canon, but I've read basically zero and couldn't say.
This reminds me of two things one is short and pithy. Hobbes said their were two sciences were we make everything and so admit of demonstrative knowledge (sure inference from first principles):
Geometry therefore is demonstrable for the lines and figures from which we reason are drawn and described by ourselves and civil philosophy is demonstrable because we make the commonwealth ourselves. (EW VII.184)
(This quote is apparently from Six Lessons to the Professors of Mathematiques.)
This overestimates the regularity or surety of demonstration we can achieve in civil philosophy.
More substantively it reminds me of the distinction made famous by Einstein between two types of scientific theories. Constructive theories as in statistical mechanics were a theory is developed from a simple hypothesis about some underlying mechanism. So this seems roughly similar to your top down concept. To this Einstein contrasts theories of principle like Einstein's own theory relativity or thermodynamics, which simply mathematically codify empirically observed generalizations. So this would roughly correspond to your bottom up concepts.
Einstein elaborates this distinction to elucidate issues with relativity in a popular account from 1919, "Time, Space, and Gravitation"
https://en.wikisource.org/wiki/Time,_Space,_and_Gravitation
The basic distinction was already made by the 19th century Scottish physicist William John Macquorn Rankine in 1855 and had been discussed by various people including Pierre Duhem, which is where I assume Einstein got it from. Rankine contrasts hypothetical method of theorizing (that posit some unsee mechanism) against abstractive method (that identifies sets of things that have been observed by the senses to have common properties). Instead of hypothetical Morris Cohen and Ernest Nagel say physical and instead of abstractive they say mathematical.
Rankine makes his distinction in this paper: https://en.wikisource.org/wiki/Outlines_of_the_Science_of_Energetics
I learned of Rankine from happening to read a little of Cohen and Nagel's 1936 book (they don't actually connect Rankine and Einstein, but its clear the exposition must be related, I guess Duhem is the likely connection because Einstein famously read some Duhem, Einstein could easily have read Rankine himself I spose):
https://archive.org/details/in.ernet.dli.2015.499820/page/n3/mode/2up
An Introduction to Logic and Scientific Method
I have expanded this reference to Einstein and others thoughts about scientific theories into a post. This is only tangentially related to the initial post, but this seems like a good place to link to it anyway. https://substack.com/home/post/p-173908686
Excellent piece! I agree with pretty much all of it. I share your skepticism of the opaque/transparent concept distinction and find the claims about being "super justified" about the nature of your own internal states extremely dubious, to say the least. Though for slightly different reasons. Our access to our internal states is just as theory-laden, fallible and subject to revision as any putative belief about the external world, Cartesian dogmatism notwithstanding. I'm inclined to think there isn't actually such thing as concepts that 'reveal essences' at all, and whether you count as knowing what a given concept refers to, depends on the (largely empirical) background knowledge you have. Though, the points you make here are also interesting and make me want to read into ordinary language philosophy which I've held off for a while!
This is a precise piece and I suspect (though I hold this tentatively) that the top-down/bottom-up distinction you’ve drawn may be doing even more than you’ve said here.
It may actually be the case that there are two distinct populations of language users who differ in what functions as a token versus a type in their processing. For one population **world-token processors** tokens are real-world phenomenological referents and words are then transparent pointers to felt experience. For this population ‘Types’ are fuzzy categories, bell curves of grouped, similar phenomenology.
For a second population **word-token processors** tokens are the words themselves. Precise, discrete, sharp-edged lexemes. Types for us are patterns of use, distributional roles, operators rather than felt categories. The phenomenological undergirding that justifies word usage for the first population is. I suspect, largely unavailable as a first-person justification for us.
And yes I suspect I’m in the second population and perhaps you may be also. And I further suspect we may be in the minority.
What makes this potentially interesting rather than merely taxonomic is a layering observation. There may be something like three consecutive layers: phenomenological experience, words that point at that experience, and the operator structure of how those words behave grammatically. World-token processors have transparent access to the first two layers — the felt experience that then generates the word that points to it. Word-token processors are, I suspect, largely out of phase with the first layer and find themselves instead needing a process, maybe even a hack to know when to use the words. This process or operator later is the third and we use it precisely because the first layer isn’t generating the meaning.
Being out of phase in this way may put a small minority in the position of seeing the operator structure clearly while finding the phenomenological grounding that most people take as obvious somewhat opaque. The parsimonious account isn’t that one population is reasoning correctly and the other isn’t. It’s that each population has transparent access to different layers of the same structure.
If this is right, your top-down/bottom-up distinction may be tracking something real not just about concepts but about the processors using them. I’d be curious whether any of this resonates.
"I don’t think there are concepts that (for some mysterious reason) are a bit voyeuristic and strip down so we can see all their tidbits, while others are more coy and waltz about clothed, revealing only an ankle here or a sultry gaze over there."
I believe that would make the concepts exhibitionists. We're the voyeurs, hoping to catch a glimpse of some top-down/bottom-up action.
to the repository!
Yeah this discussion reminds me of other things that you and I have spoken about before. The only thing I disagree with is when you say that "analytic philosophy goes astray" because of its failures to recognize these sorts of things. I would say that analytic philosophy EXISTS primarily because of these failures, and whether they're mistakes or not, they are very important parts of the basement-level presuppositions that make analytic philosophy what it is. In other words, if you don't share these presuppositions about language and thought, there is surprisingly little for you to do in mainstream analytic philosophy, because most of it depends on the kind of things you were talking about.
I think in the past you and I have called it a "residual Platonism" about concepts. As far as I can tell, mainstream analytic philosophy is thoroughly Platonist about concepts, in the sense that it takes it for granted that concepts have some special "essence" that we can "discover," and that grasping the content of a concept is a true and significantly meaningful cognitive achievement, kind of like popping your head outside Plato's cave and seeing by the light of the sun. It's like a genuine DISCOVERY to find out something about a particular concept, instead of, as you said, a simple recognition of what we say or stipulate. Inasmuch as analytic philosophy has articles of faith, this bit about concepts is one of them.
I remember having a discussion about this in one of my graduate seminars. We were reading an ethics paper with various thought experiments that involved the concept of a "killing," and various intuitions pushed people to say that one thing counted as a "killing" and another didn't. And it's like the professor felt that there was truly something meaningful to be discovered, the essence of the concept, if we could only sort through all the different factors. But in reality, a "killing" is whatever we want it to be, and so if we wanted a ham sandwich to count as a killing, it would count as a killing. There's nothing further to discover about concepts than that a lot of the time.
I think this is a little overstating the case against analytic philosophy. Analyzing even purely stipulative concepts can result in genuine discoveries in just the same way that mathematics can. It's often not clear at all what a collection of formal stipulations implies.
There is, however, a big problem with the use of thought experiments to discover the content of a concept.
I guess if by "discovery" you mean something like "pragmatic insight into how we could construct a concept differently in order to better serve our ends," then I would agree. I can't see how we could ever get more than that by analysis.
I'm thinking of how carrying out deductions based on formal definitions can prove statements we didn't know those formal definitions committed us to. If we stipulated different axiomatizations of set theory, we wouldn't thereby immediately know what downstream theorems were entailed. We'd have to reason through it.
Oh ok, that's different from what I was thinking, and it seems right.
Great write up. "Transparent concepts" are seriously problematic. I especially like the note about slipping between the ostensive definition and a stipulated definition. The formalized concept possesses its features essentially because it's a formalization, and it *does not* thereby follow that the ostensive concept actually has those features.
If I can't doubt that pain is bad, it doesn't follow that *this* pain is certainly bad--it follows that *if this is indeed pain*, then it's bad. But perhaps what I take to be pain isn't pain, just as you say.
This strikes me as a variant of the fixed points argument for moral realism. Seeing as that argument is extremely weak, this one doesn’t do much better. I’ve never been a big fan of these stipulative moves that realists make, and they certainly have little motivational force.
In the original post, the author claims
“When I stub my toe, I immediately recognise the badness of my pain.” I interpret this to mean something like “I have perception, in the same way as hearing, seeing, etc that pain is (normatively) bad.” I also find moral perception arguments to be pretty weak, as they venture into the realm of empirics, where I’ve seen no evidence to support that people literally and directly perceive moral properties.
Excellent. I think the TD and BU concepts distinction is thoroughly explanatory.