No record in English (and I’m using English as a shorthand for any human language) can ever be good enough. English is not a technology for transmitting information.
English is a compression format, and a very lossy and somewhat inaccurate compression format at that. But it has a stupendously high compression rate and compression algorithms with reasonable running speeds on specially adapted hardware (i.e. brains), so for the particular purposes of human communication English is a pretty decent option.
I own a t-shirt with this graphic printed on it. If you possess a mostly correct compression algorithm (that is, you speak modern English), the ~5kb of data on that shirt contains sufficient information to reproduce ~30 major scientific or technological discoveries. I don’t know exactly how much space you could fit that information into if you encoded it in a way that wasn’t very heavily optimized for very specific types of human brains, but I suspect it’s many orders of magnitude greater than 5kb.
On the surface this seems like it could be an argument for reproducing a specific human from their preserved written material, what with the amazing information density of English. But using a standard English decompression algorithm to analyze what you’ve written is worthless, because we’re not trying to recreate the meaning of what you’ve written. We’re trying to recreate the compression algorithm used to create your writings, which would be approximately isomorphic with your brain. But because the English data format is lossy and imprecise, reconstructing that algorithm from only its output is impossible.
If you could preserve a copy of the decompressed version of what you were trying to write along with your writings, that might be enough to reverse-engineer your brain(’s compression algorithm). But I don’t think that’s possible for any human, much less most of them.
How is your description of English as a compression format different from the idea of the detached lever, where one puts the characters a p p l e into a computer and hopes it will have crunchy, juicy properties?
I believe I speak Modern English, and could probably look for wavy hands on sticks penicillin mold or coil wires around magnets, but how does “atoms can be split” help me reproduce a major scientific/engineering discovery? It’s not a compressed instruction, it’s a teachers password I can say to other people who know that atoms can be split so we can be comfortably “scientific” in each other’s presence. I don’t know what it means in terms of equations, machinery, or testable predictions—and more to the point, - I still don’t know what it means after reading the t-shirt.
I could probably grope about in a corpse and find a heart or a lung, but how do I tell when I have a pancreas instead of a phlogistondix? And which bit of it is the pancreatic duct? And how do I tell if the fluid that comes out of the unknown lump of creature that I have is insulin? Or after injection, how to tell if it’s working?
The only constrained anticipation I have for ‘insulin’ is that it helps diabetic people—although I now note that I have no real idea what ‘diabetic people’ means in medical terms or how, if I were thrust back in time, I would be able to reliably identify them.
I suggest that t-shirt is not a compressed guide, it’s a memory aid for people who already know the details behind it and who could, if their memory was entirely under their command, manage exactly the same without it.
English is a compression algorithm, but most of the information required for that algorithm is stored in your brain. Your brain hears the word “apple” and expands it to represent everything that you know about apples. If your brain can’t expand “pancreas” as far, that is a characteristic of your brain and not the word.
As is true of software compression algorithms, the purpose of your brain’s compression algorithm is to allow you to shrink the size of your knowledge and messages, at the cost of computing time and accuracy.
The only constrained anticipation I have for ‘insulin’ is that it helps diabetic people—although I now note that I have no real idea what ‘diabetic people’ means in medical terms or how, if I were thrust back in time, I would be able to reliably identify them.
I suggest that t-shirt is not a compressed guide, it’s a memory aid for people who already know the details behind it and who could, if their memory was entirely under their command, manage exactly the same without it.
But these terms don’t exist in complete isolation. Say for example I’m sent back to 1850. Then I don’t know what the different parts of a pancreas look like, but doctors will know. So I can bootstrap my knowledge based on that (and presumably they know what a diabetic is and how to recognize them). Some of these (like using quartz crystals to make clocks) are difficult due to infrastructural problems, but most of them have large amounts of associated ideas that connect to the terms.
By analogy with the issue being discussed, the terms being used don’t function completely as detached levers, since when we have a written record of you saying “I like to eat apples but not oranges” we have a specific idea of what “apple” means.
Are you saying that once you have a written record of me mentioning apples, then you can talk to me about ‘apples’ with no explanation, but before that you would have to talk to me about ‘apples (which are …)’ with an explanation?
Are you saying that once you have a written record of me mentioning apples, then you can talk to me about ‘apples’ with no explanation, but before that you would have to talk to me about ‘apples (which are …)’ with an explanation?
Hmm, ok. That can’t be right when phrased that way. So something is wrong with my notions. It may be that the point about time-travel holds but generalizing it to the lever issue fails.
I believe that language is for communicating the shared part of experience, or sometimes for creating the illusion of shared experience. Whatever is unique about a person’s experience is going to get lost if you try to communicate it through language.
Ok, that’s maybe a little too harsh-sounding. I think some people are relatively similar to each other, so that language can resonate rfairly well between them.
Still, I believe in tacit knowledge. And even if a skillful person can find words for some of it—turn the bike wheel towards the direction you’re falling is sound advice, but how would you convey exactly what it’s like to be you riding a bike on a particular day, or what it’s like to know how to ride a bike before you have words for it?
No record in English (and I’m using English as a shorthand for any human language) can ever be good enough. English is not a technology for transmitting information.
English is a compression format, and a very lossy and somewhat inaccurate compression format at that. But it has a stupendously high compression rate and compression algorithms with reasonable running speeds on specially adapted hardware (i.e. brains), so for the particular purposes of human communication English is a pretty decent option.
I own a t-shirt with this graphic printed on it. If you possess a mostly correct compression algorithm (that is, you speak modern English), the ~5kb of data on that shirt contains sufficient information to reproduce ~30 major scientific or technological discoveries. I don’t know exactly how much space you could fit that information into if you encoded it in a way that wasn’t very heavily optimized for very specific types of human brains, but I suspect it’s many orders of magnitude greater than 5kb.
On the surface this seems like it could be an argument for reproducing a specific human from their preserved written material, what with the amazing information density of English. But using a standard English decompression algorithm to analyze what you’ve written is worthless, because we’re not trying to recreate the meaning of what you’ve written. We’re trying to recreate the compression algorithm used to create your writings, which would be approximately isomorphic with your brain. But because the English data format is lossy and imprecise, reconstructing that algorithm from only its output is impossible.
If you could preserve a copy of the decompressed version of what you were trying to write along with your writings, that might be enough to reverse-engineer your brain(’s compression algorithm). But I don’t think that’s possible for any human, much less most of them.
How is your description of English as a compression format different from the idea of the detached lever, where one puts the characters a p p l e into a computer and hopes it will have crunchy, juicy properties?
I believe I speak Modern English, and could probably look for wavy hands on sticks penicillin mold or coil wires around magnets, but how does “atoms can be split” help me reproduce a major scientific/engineering discovery? It’s not a compressed instruction, it’s a teachers password I can say to other people who know that atoms can be split so we can be comfortably “scientific” in each other’s presence. I don’t know what it means in terms of equations, machinery, or testable predictions—and more to the point, - I still don’t know what it means after reading the t-shirt.
I could probably grope about in a corpse and find a heart or a lung, but how do I tell when I have a pancreas instead of a phlogistondix? And which bit of it is the pancreatic duct? And how do I tell if the fluid that comes out of the unknown lump of creature that I have is insulin? Or after injection, how to tell if it’s working?
The only constrained anticipation I have for ‘insulin’ is that it helps diabetic people—although I now note that I have no real idea what ‘diabetic people’ means in medical terms or how, if I were thrust back in time, I would be able to reliably identify them.
I suggest that t-shirt is not a compressed guide, it’s a memory aid for people who already know the details behind it and who could, if their memory was entirely under their command, manage exactly the same without it.
I agree with JoshuaZ, but would add:
English is a compression algorithm, but most of the information required for that algorithm is stored in your brain. Your brain hears the word “apple” and expands it to represent everything that you know about apples. If your brain can’t expand “pancreas” as far, that is a characteristic of your brain and not the word.
As is true of software compression algorithms, the purpose of your brain’s compression algorithm is to allow you to shrink the size of your knowledge and messages, at the cost of computing time and accuracy.
But these terms don’t exist in complete isolation. Say for example I’m sent back to 1850. Then I don’t know what the different parts of a pancreas look like, but doctors will know. So I can bootstrap my knowledge based on that (and presumably they know what a diabetic is and how to recognize them). Some of these (like using quartz crystals to make clocks) are difficult due to infrastructural problems, but most of them have large amounts of associated ideas that connect to the terms.
By analogy with the issue being discussed, the terms being used don’t function completely as detached levers, since when we have a written record of you saying “I like to eat apples but not oranges” we have a specific idea of what “apple” means.
Are you saying that once you have a written record of me mentioning apples, then you can talk to me about ‘apples’ with no explanation, but before that you would have to talk to me about ‘apples (which are …)’ with an explanation?
Hmm, ok. That can’t be right when phrased that way. So something is wrong with my notions. It may be that the point about time-travel holds but generalizing it to the lever issue fails.
I believe that language is for communicating the shared part of experience, or sometimes for creating the illusion of shared experience. Whatever is unique about a person’s experience is going to get lost if you try to communicate it through language.
Ok, that’s maybe a little too harsh-sounding. I think some people are relatively similar to each other, so that language can resonate rfairly well between them.
Still, I believe in tacit knowledge. And even if a skillful person can find words for some of it—turn the bike wheel towards the direction you’re falling is sound advice, but how would you convey exactly what it’s like to be you riding a bike on a particular day, or what it’s like to know how to ride a bike before you have words for it?