I am not convinced by the premise, mostly because I think that the tradeoff between practice and knowledge is illusory.
Tacit knowledge is knowledge that cannot be captured through words alone.
Most knowledge is not tacit; we can and do use words to transmit the underlying concepts. Indeed, the author is using words alone to communicate the idea of tacit knowledge. In associated articles, the author contradicts their own definition. In The Mental Model FAQ, the author instead defines tacit knowledge in terms of mental representations, and explains that tacit knowledge is implicit in the mental representation of the knower:
Mental representations are difficult — if not impossible! — to communicate explicitly. They are tacit in nature.
So, yes, mental models are incredibly useful. If nothing else, they tell you that the valuable bits of expertise is not what is communicable, but what is tacit and stuck inside the practitioner’s head.
But there also exist mental models that are nearly impossible to communicate. We call this tacit knowledge.
Polanyi argued that there is knowledge that cannot be adequately articulated by verbal means, and asserted that most of human knowledge is rooted in tacit knowing.
This is not quite the same as the top-posted article, though. Is tacit knowledge unable to be expressed with words, or difficult to express with words? There is a massive gulf between these two situations, and the gulf is wide and deep enough that, for anything unable to be expressed with words, I expect a proof along the lines of Berry’s paradox. Interesting numbers are not definable numbers.
Really, to be frank, what I think that the author is so close to discovering is that words have no meanings. Since words have no meanings, there isn’t any actual semantic content transferred by an explanation. Therefore, regardless of the number of words used, nobody can really teach anybody how to do anything; words might be a guide, but every learner must discover every fact for themselves, and cannot rely on words to show them how it might be done, because words cannot carry semantics, because words have no meanings. Through this viewpoint, all knowledge is tacit, and all written words are merely reflections of how people feel about knowing things.
Really, to be frank, what I think that the author is so close to discovering is that words have no meanings.
I get where you are coming from here and would agree that in the end, everyone has to learn something on their own. However, I think a bigger issue might be confusing experience and rhetorical ability.
I have often had this debate with colleagues, whether we should look for people with innate ability, a good ability to learn, or that can follow instructions well and which of those factors made us successful. I do believe some people learn quicker than others, but the crux of his argument seems to not be the speed at which someone picks something up, but the ability to render it into words.
I am often guilty of speaking like his example “Well, do X. Except when you see Y, then do Z, because A. And if you see B, then do P. But if you see A and C but not B, then do Q, because reason D. And then there are weird situations where you do Z but then see thing C emerge, then you should switch to Q.” But this is not because I have some tacit knowledge the listener does not, it is because I cannot appropriately articulate the underlying heuristics that need to be learned. Often times, if we thought about it more, I’ll bet the example ends up much more refined, closer to codified steps, but the speaker never had to think about the underlying rules.
To illustrate what I mean, I would use the writer’s example of teaching a kid to ride a bike. I would have no idea how to approach that, but the writer was able to distill it down to a step by step guide. He thought about it and found the appropriate words.
So overall, I’d agree he misses the mark, but not as much because words have no meaning as because he is assuming someone who can perform well in their field also has the rhetorical and logical capacity to render their experience into appropriate language.
To illustrate what I mean, I would use the writer’s example of teaching a kid to ride a bike. I would have no idea how to approach that, but the writer was able to distill it down to a step by step guide. He thought about it and found the appropriate words.
I disagree. He distilled riding a bike to a step-by-step guide to learning. He was unable to distill it to a step-by-step guide to riding for someone who hadn’t learned.
Fair point, I chose a bad example. What I meant was if presented today with the challenge of telling someone how to teach a child to ride a bike, I wouldn’t know how to begin. I have neither the words, nor the experience. The writer had the experience and came up with decent enough words that I could begin to gain the experience, had I a child and a small bike. But, you are right, he has not expressed the right words for riding in here. Does that mean they can’t exist?
I don’t think it means they can’t. The author cited Boyd as an example of one case where someone did figure out how to put into words a thing that had been widely considered unexplainable.
One point I took away from the piece is that it’s more efficient to set about gaining the “unexplainable” knowledge through guided experience (as in the bike riding example) as opposed to try and brute force explain it or read up on it.
(And if you ever find yourself wanting to teach a small child to ride a bicycle, the “balance bike” is an invention that significantly accelerates the learning the author explained in that example. Source: I’m not much good at bike riding myself, damn sure can’t explain how, and taught two small people how to ride by helping them get the feel on “balance bikes” first.)
Are we self-aware enough articulate the underlying heuristics? Humans are biased in lots of different ways and I would assume that all our heuristics are also biased in ways we are not aware of.
One symptom is that seem to forget stuff: “You saw A and C but not B, so you tried Q but it did not help. Oh, did you also check E? In that case Q is of course wrong and you should do R instead.” Did we actually forget it or do we rationalize heuristics below our awareness? If we rationalize it then we can never know if the explicit heuristics are complete or if more special cases are lurking. That seems to match the futility of expert systems where a lot of effort was made to codify such decision making. For some reason there was always something only the humans could do though.
You really think there is a “massive gulf” between theoretically impossible and practically impossible? Such pedantry! No wonder the poor guy had to write this (excellent, IMO) essay.
Take any Linguistics 101 and you’ll learn very quickly how much tacit knowledge you needed to have just to express or understand a grammatical sentence in the first place.
Isn’t “tacit knowledge is knowledge that cannot be captured through words alone” like Berry’s paradox “the smallest positive integer not definable in under sixty letters”? Of course, we can now discuss what “captured” means just like Wikipedia describes the discussion about “definable”.
The concept of ‘tacit knowledge’ that the author is describing sounds like what I would call ‘intuition’ – a type of unconscious skill gained through experience.
I am not persuaded that this is the right frame to argue back and forth on. A few points.
Is he arguing expertise is impossible to communicate, or just very hard? Is he arguing that the map is incommesurable with the territory? Knowledge acquisition is an important problem, but I think he’s conflating serious philosophical problems with a pedagogical approach. Also, he is conflating different classes of skills - motor and intellectual, and that’s not a good way to tackle a problem either.
From a logical analysis, he asserts his conclusion, and thus undermines what he’s trying to get at.
I am not convinced by the premise, mostly because I think that the tradeoff between practice and knowledge is illusory.
Most knowledge is not tacit; we can and do use words to transmit the underlying concepts. Indeed, the author is using words alone to communicate the idea of tacit knowledge. In associated articles, the author contradicts their own definition. In The Mental Model FAQ, the author instead defines tacit knowledge in terms of mental representations, and explains that tacit knowledge is implicit in the mental representation of the knower:
In The Mental Model Fallacy, the author uses “tacit” as in the top-posted article:
This is not quite the same as the top-posted article, though. Is tacit knowledge unable to be expressed with words, or difficult to express with words? There is a massive gulf between these two situations, and the gulf is wide and deep enough that, for anything unable to be expressed with words, I expect a proof along the lines of Berry’s paradox. Interesting numbers are not definable numbers.
Really, to be frank, what I think that the author is so close to discovering is that words have no meanings. Since words have no meanings, there isn’t any actual semantic content transferred by an explanation. Therefore, regardless of the number of words used, nobody can really teach anybody how to do anything; words might be a guide, but every learner must discover every fact for themselves, and cannot rely on words to show them how it might be done, because words cannot carry semantics, because words have no meanings. Through this viewpoint, all knowledge is tacit, and all written words are merely reflections of how people feel about knowing things.
I get where you are coming from here and would agree that in the end, everyone has to learn something on their own. However, I think a bigger issue might be confusing experience and rhetorical ability.
I have often had this debate with colleagues, whether we should look for people with innate ability, a good ability to learn, or that can follow instructions well and which of those factors made us successful. I do believe some people learn quicker than others, but the crux of his argument seems to not be the speed at which someone picks something up, but the ability to render it into words.
I am often guilty of speaking like his example “Well, do X. Except when you see Y, then do Z, because A. And if you see B, then do P. But if you see A and C but not B, then do Q, because reason D. And then there are weird situations where you do Z but then see thing C emerge, then you should switch to Q.” But this is not because I have some tacit knowledge the listener does not, it is because I cannot appropriately articulate the underlying heuristics that need to be learned. Often times, if we thought about it more, I’ll bet the example ends up much more refined, closer to codified steps, but the speaker never had to think about the underlying rules.
To illustrate what I mean, I would use the writer’s example of teaching a kid to ride a bike. I would have no idea how to approach that, but the writer was able to distill it down to a step by step guide. He thought about it and found the appropriate words.
So overall, I’d agree he misses the mark, but not as much because words have no meaning as because he is assuming someone who can perform well in their field also has the rhetorical and logical capacity to render their experience into appropriate language.
I disagree. He distilled riding a bike to a step-by-step guide to learning. He was unable to distill it to a step-by-step guide to riding for someone who hadn’t learned.
Fair point, I chose a bad example. What I meant was if presented today with the challenge of telling someone how to teach a child to ride a bike, I wouldn’t know how to begin. I have neither the words, nor the experience. The writer had the experience and came up with decent enough words that I could begin to gain the experience, had I a child and a small bike. But, you are right, he has not expressed the right words for riding in here. Does that mean they can’t exist?
Riding a bike is a lot about trained reflexes. I’m very confident that no amount of words can instill reflexes in somebody.
Maybe a more direct connection to neurons can do that? Like the Matrix movie depicts it as Neo learns Kung Fu which is also a lot about reflexes.
I don’t think it means they can’t. The author cited Boyd as an example of one case where someone did figure out how to put into words a thing that had been widely considered unexplainable.
One point I took away from the piece is that it’s more efficient to set about gaining the “unexplainable” knowledge through guided experience (as in the bike riding example) as opposed to try and brute force explain it or read up on it.
(And if you ever find yourself wanting to teach a small child to ride a bicycle, the “balance bike” is an invention that significantly accelerates the learning the author explained in that example. Source: I’m not much good at bike riding myself, damn sure can’t explain how, and taught two small people how to ride by helping them get the feel on “balance bikes” first.)
Are we self-aware enough articulate the underlying heuristics? Humans are biased in lots of different ways and I would assume that all our heuristics are also biased in ways we are not aware of.
One symptom is that seem to forget stuff: “You saw A and C but not B, so you tried Q but it did not help. Oh, did you also check E? In that case Q is of course wrong and you should do R instead.” Did we actually forget it or do we rationalize heuristics below our awareness? If we rationalize it then we can never know if the explicit heuristics are complete or if more special cases are lurking. That seems to match the futility of expert systems where a lot of effort was made to codify such decision making. For some reason there was always something only the humans could do though.
You really think there is a “massive gulf” between theoretically impossible and practically impossible? Such pedantry! No wonder the poor guy had to write this (excellent, IMO) essay.
Take any Linguistics 101 and you’ll learn very quickly how much tacit knowledge you needed to have just to express or understand a grammatical sentence in the first place.
Isn’t “tacit knowledge is knowledge that cannot be captured through words alone” like Berry’s paradox “the smallest positive integer not definable in under sixty letters”? Of course, we can now discuss what “captured” means just like Wikipedia describes the discussion about “definable”.
More on tacit knowledge featured recently:
https://lobste.rs/s/3s309w/programming_as_theory_building
The concept of ‘tacit knowledge’ that the author is describing sounds like what I would call ‘intuition’ – a type of unconscious skill gained through experience.
I am not persuaded that this is the right frame to argue back and forth on. A few points.
Is he arguing expertise is impossible to communicate, or just very hard? Is he arguing that the map is incommesurable with the territory? Knowledge acquisition is an important problem, but I think he’s conflating serious philosophical problems with a pedagogical approach. Also, he is conflating different classes of skills - motor and intellectual, and that’s not a good way to tackle a problem either.
From a logical analysis, he asserts his conclusion, and thus undermines what he’s trying to get at.
At the beginning, the author states:
And at the end:
pass