one has to look at the purpose of an idea instead of just the raw data
Oh, I agree. It’s just that you were very insistent about drawing the line between unfalsifiable philosophy and other empirically-falsifiable stuff and here you’re coming back into the real-life problems realm where things are definitely testable and falsifiable. I’m all for it, but there are consequences.
So you have to talk about your interpretation, unless the other person can guess it.
Sure, but that’s not an intellectual debate. If someone asks how to start a fire and I explain how you arrange kindling, get a flint and a steel, etc. there is no debate—I’m just transferring information.
the philosophy issues are always prior to the other issues
Not necessarily. If you put your hand into a fire, you will get a burn—that’s easy to learn (and small kids learn it fast). Which philosophy issues are prior to that learning?
Please stop the sarcasm
No can do. But tell you what, the fewer silly things you say, the less often you will encounter overt sarcasm :-)
in order to solve epistemology problems
Which problems you can’t solve otherwise?
for any given issue you decide yes to a particular thing and no to its rivals
There are lot of issues with continuous (real number) decisions. Let’s say you’re deciding how much money to put into your retirement fund this year and the reasonable range is between $10K and $20K. You are not going to treat $14,999 and $15,000 as separate solutions, are you?
breakpoints like this come up all over the place
Sure they do, but not always. And your approach requires them.
the decision should itself by judged in a binary way and be non-refuted
I still don’t see the need for these rather severe limitations. You want to deal with reality as if it consists of discrete, well-delineated chunks and, well, it just doesn’t. I understand that you can impose thresholds and breakpoints any time you wish, but they are artifacts and if your method requires them, it’s a drawback.
then do whatever maximizes it
Yes, but you typically have an explore-or-exploit problem. You need to spend resources to look for a better optimum, at each point in time you have some probability of improving your maximum, but there are costs and they grow. At which point do you stop expending resources to look for a better solution?
It’s just that you were very insistent about drawing the line between unfalsifiable philosophy and other empirically-falsifiable stuff
if you have an empirical argument to make, that’s fine. but i don’t think i’m required to provide evidence for my philosophical claims. (btw i criticize the standard burden of proof idea in Yes or No Philosophy. in short, if you can’t criticize an idea then it’s non-refuted and demanding some sort of burden of proof is not a criticism since lack of proof doesn’t prevent an idea from solving a problem.)
in order to solve epistemology problems
Which problems you can’t solve otherwise?
the problem of induction. problems about how to evaluate arguments (how do you score the strength of an argument? and what difference does it really make if one scores higher than another? either something points out why a solution doesn’t work or it doesn’t. unless you specifically try to specify non-binary problems. but that doesn’t really work. you can specify a set of solutions are all equal. ok then either pick any one of them if you’re satisfied, or else solve some other more precise problem that differentiates. you can also specify that higher scoring solutions on some metric are better, but then you just pick the highest scoring one, so you get a single solution or maybe a tie again. and whether you’ve chosen a correct solution given the problem specification, or not, is binary.) and various problems about how you decide what metrics to use (the solution to that being binary arguments about what metrics to use – or in many cases don’t use a metric. metrics are overrated but useful sometimes.)
Yes, but you typically have an explore-or-exploit problem. You need to spend resources to look for a better optimum, at each point in time you have some probability of improving your maximum, but there are costs and they grow. At which point do you stop expending resources to look for a better solution?
Yes so then you guess what to do and criticize your guesses. Or, if you wish, define a metric with positive points for a higher score and negative points for resources spent (after you guess-and-criticize to figure out how to put the positive score and all the different types of resources into the same units) and then guess how to maximize that (e.g. define a metric about resources allocated to getting a higher score on the first metric, spend that much resources, and then use the highest scoring solution.
multi-factor metrics don’t work as well as people think, but are ok sometimes (but you have to make a binary judgement about whether to use a particular metric for a particular situation, or not – so the binary judgement is prior and governs the use of the metric). here’s a good article about issues with them:
scoring systems are overrated but are allowable in binary epistemology given that their use is governed by binary judgements (should I proceed by doing the thing that scores the highest on this metric? make critical arguments about that and make a binary judgement. so the binary judgement is prior but then things like metrics and statistics are allowable as secondary things which are sometimes quite useful.)
You are not going to treat $14,999 and $15,000 as separate solutions, are you?
depends how precise the problem or context says to be. (or bigger picture, it depends how precise is worth the resources to be – which you should either specify in the problem or consider part of the context.)
if you don’t care about single dollar level of precision (cuz you want to save resources like effort to deal with details), just e.g. specify in the problem that you only care about increments of $500 or that (to save problem solving resources like time) you just want to use the first acceptable solution you come up with that you determine meet some standards of good enough (these are no longer strictly single variable maximization problems).
breakpoints like this come up all over the place
Sure they do, but not always. And your approach requires them.
they aren’t required, you can specify the problem however you want (subject to criticism) so you it makes clear what is a solution or not (or a set of tied solutions you’re indifferent btwn which you can then tiebreak arbitrarily if you have no criticism of doing it arbitrarily).
if the problem specifies that some solutions are better than others (not my preferred way to specify problems – i think it’s epistemologically misleading), then when you act you should pick one of the solutions in the highest tier you have a solution in, and reject the others. whether this method (pick a highest tier solution) is correct, and whether you’ve used it in this case, are both binary issues open to criticism.
At which point do you stop expending resources to look for a better solution?
when you guess it’s best to stop and your guess is non-refuted and the guess to continue looking is refuted. (you may, if you want to, define some stopping metric and make a subject-to-criticism binary yes-or-no judgement about whether to use that stopping metric.)
the philosophy issues are always prior to the other issues
Not necessarily. If you put your hand into a fire, you will get a burn—that’s easy to learn (and small kids learn it fast). Which philosophy issues are prior to that learning?
i think small kids do guesses and criticism, and use methods of learning (what I would call philosophical methods), even if they can’t state those methods in English. i also think ppl who have never studied philosophy use philosophy methods, which they picked up from their culture here and there, even if they don’t consciously understand themselves or know the names of the things they’re doing. and to the extent ppl learn, i think it’s guesses and criticism in some form, since that’s the only known method of learning (at a low level, it’s evolution – the only known solution to the problem of where the appearance of design comes from – saying it comes from “intelligence” is like attributing it to God or an intelligent designer – it doesn’t tell you how god/intelligence does it. my answer to that is, at a low level, evolution. layers of abstraction are built on top of that so it looks more varied at a higher level.).
i don’t think i’m required to provide evidence for my philosophical claims
It depends on what do you want to do with them. If all you want to do is keep them on a shelf and once in a while take them out, dust them, and admire them, then no, you don’t. On the other hand, if you want to persuade someone to change their mind, evidence might be useful. And if you want other people to take action based on your claims’, ahem, implications, evidence might even be necessary.
the problem of induction. problems about how to evaluate arguments
It seems that the root of these problems is your insistence that truth is a binary category. If you are forced to operate with single-bit values and have to convert every continuous function into a step one, well, sure you will have problems.
The thread seem to be losing shape, so let’s do a bit of a summary. As far as I can see, the core differences between us are:
You think truth (and arguments) are binary, I think both have continuous values;
You think intellectual debates are primary and empirical testing is secondary, I think the reverse;
do you think there’s a clear, decisive mistake in something i’m saying?
I would probably classify it as suboptimal. It’s not a “clear, decisive mistake” to see only black and white—but it limits you.
can you specify how you think induction works?
In the usual way: additional data points increase the probability of the hypothesis being correct, however their influence tends to rapidly decline to zero and they can’t lift the probability over the asymptote (which is usually less than 1). Induction doesn’t prove anything, but then in my system nothing proves anything.
What you said in the previous message is messy and doesn’t seem to be terribly impactful. Talking about how you can define a loss function or how you can convert scores to a yes/no metric is secondary and tertiary to the core disagreements we have.
the probability of which hypotheses being correct, how much?
For a given problem I would have a set of hypotheses under consideration. A new data point might kill some of them (in the Popperian fashion) or might spawn new ones. Those which survive—all of them—gain some probability. How much, it depends. No simple universal rule.
how do you differentiate hypotheses which do not contradict any of the data?
For which purpose and in which context? I might not need to differentiate them.
Occam’s razor is a common heuristic, though, of course, it is NOT a guide to whether a particular theory is correct or not.
Do all the non-contradicted-by-evidence ideas gain equal probability (so they are always tied and i don’t see the point of the “probabilities”), or differential probability?
EDIT: I’m guessing your answer is you start them with different amounts of probability. after that they gain different amounts accordingly (e.g. the one at 90% gains less from the same evidence than the one at 10%). but the ordering (by amount of probability) always stays the same as how it started, apart from when something is dropped to 0% by contradicting evidence. is that it? or do you have a way (which is part of induction, not critical argument?) to say “evidence X neither contradicts ideas Y nor Z, but fits Y better than Z”?
Different hypotheses (= models) can gain different amounts of probability. They can start with different amounts of probability, too, of course.
to say “evidence X neither contradicts ideas Y nor Z, but fits Y better than Z”?
Of course. That’s basically how all statistics work.
Say, if I have two hypotheses that the true value of X is either 5 or 10, but I can only get noisy estimates, a measurement of 8.7 will add more probability to the “10” hypothesis than to the “5″ hypothesis.
They get identical probabilities—if their prior probabilities were equal.
If (as is the general practice around these parts) you give a markedly bigger prior probability to simpler hypotheses, then you will strongly prefer the simpler idea. (Here “simpler” means something like “when turned into a completely explicit computer program, has shorter source code”. Of course your choice of language matters a bit, but unless you make wilfully perverse choices this will seldom be what decides which idea is simpler.)
In so far as the world turns out to be made of simply-behaving things with complex emergent behaviours, a preference for simplicity will favour ideas expressed in terms of those simply-behaving things (or perhaps other things essentially equivalent to them) and therefore more-explanatory ideas. (It is at least partly the fact that the world seems so far to be made of simply-behaving things with complex emergent behaviours that makes explanations so valuable.)
I do, but more or less only to the extent that they will make potential different predictions. If two models are in principle incapable of making different predictions, I don’t see why should I care.
so e.g. you don’t care if trees exist or not? you think people should stop thinking in terms of trees and stick to empirical predictions only, dropping any kind of non-empricial modeling like the concept of a tree?
Isn’t it convenient that I don’t have to care about these infinitely many theories?
why not?
Since there is an infinity of them, I bet you can’t marshal critical arguments against ALL of them :-P
you can criticize categories, e.g. all ideas with feature X.
I think you’re getting confused between actual trees and the abstract concept of a tree.
i don’t think so. you can’t observe entities. you have to interpret what entities there are (or not – as you advocated by saying only prediction matters)
you can criticize categories, e.g. all ideas with feature X
How can you know that every single theory in that infinity has feature X? or belongs to the same category?
you can’t observe entities
My nervous system makes perfectly good entities out of my sensory stream. Moreover, a rat’s nervous system also makes perfectly good entities out if its sensory stream regardless of the fact that the rat has never heard of epistemology and is not very philosophically literate.
or not
Or not? Prediction matters, but entities are an awfully convenient way to make predictions.
Oh, I agree. It’s just that you were very insistent about drawing the line between unfalsifiable philosophy and other empirically-falsifiable stuff and here you’re coming back into the real-life problems realm where things are definitely testable and falsifiable. I’m all for it, but there are consequences.
Sure, but that’s not an intellectual debate. If someone asks how to start a fire and I explain how you arrange kindling, get a flint and a steel, etc. there is no debate—I’m just transferring information.
Not necessarily. If you put your hand into a fire, you will get a burn—that’s easy to learn (and small kids learn it fast). Which philosophy issues are prior to that learning?
No can do. But tell you what, the fewer silly things you say, the less often you will encounter overt sarcasm :-)
Which problems you can’t solve otherwise?
There are lot of issues with continuous (real number) decisions. Let’s say you’re deciding how much money to put into your retirement fund this year and the reasonable range is between $10K and $20K. You are not going to treat $14,999 and $15,000 as separate solutions, are you?
Sure they do, but not always. And your approach requires them.
I still don’t see the need for these rather severe limitations. You want to deal with reality as if it consists of discrete, well-delineated chunks and, well, it just doesn’t. I understand that you can impose thresholds and breakpoints any time you wish, but they are artifacts and if your method requires them, it’s a drawback.
Yes, but you typically have an explore-or-exploit problem. You need to spend resources to look for a better optimum, at each point in time you have some probability of improving your maximum, but there are costs and they grow. At which point do you stop expending resources to look for a better solution?
if you have an empirical argument to make, that’s fine. but i don’t think i’m required to provide evidence for my philosophical claims. (btw i criticize the standard burden of proof idea in Yes or No Philosophy. in short, if you can’t criticize an idea then it’s non-refuted and demanding some sort of burden of proof is not a criticism since lack of proof doesn’t prevent an idea from solving a problem.)
the problem of induction. problems about how to evaluate arguments (how do you score the strength of an argument? and what difference does it really make if one scores higher than another? either something points out why a solution doesn’t work or it doesn’t. unless you specifically try to specify non-binary problems. but that doesn’t really work. you can specify a set of solutions are all equal. ok then either pick any one of them if you’re satisfied, or else solve some other more precise problem that differentiates. you can also specify that higher scoring solutions on some metric are better, but then you just pick the highest scoring one, so you get a single solution or maybe a tie again. and whether you’ve chosen a correct solution given the problem specification, or not, is binary.) and various problems about how you decide what metrics to use (the solution to that being binary arguments about what metrics to use – or in many cases don’t use a metric. metrics are overrated but useful sometimes.)
Yes so then you guess what to do and criticize your guesses. Or, if you wish, define a metric with positive points for a higher score and negative points for resources spent (after you guess-and-criticize to figure out how to put the positive score and all the different types of resources into the same units) and then guess how to maximize that (e.g. define a metric about resources allocated to getting a higher score on the first metric, spend that much resources, and then use the highest scoring solution.
multi-factor metrics don’t work as well as people think, but are ok sometimes (but you have to make a binary judgement about whether to use a particular metric for a particular situation, or not – so the binary judgement is prior and governs the use of the metric). here’s a good article about issues with them:
https://www.newyorker.com/magazine/2011/02/14/the-order-of-things
scoring systems are overrated but are allowable in binary epistemology given that their use is governed by binary judgements (should I proceed by doing the thing that scores the highest on this metric? make critical arguments about that and make a binary judgement. so the binary judgement is prior but then things like metrics and statistics are allowable as secondary things which are sometimes quite useful.)
depends how precise the problem or context says to be. (or bigger picture, it depends how precise is worth the resources to be – which you should either specify in the problem or consider part of the context.)
if you don’t care about single dollar level of precision (cuz you want to save resources like effort to deal with details), just e.g. specify in the problem that you only care about increments of $500 or that (to save problem solving resources like time) you just want to use the first acceptable solution you come up with that you determine meet some standards of good enough (these are no longer strictly single variable maximization problems).
they aren’t required, you can specify the problem however you want (subject to criticism) so you it makes clear what is a solution or not (or a set of tied solutions you’re indifferent btwn which you can then tiebreak arbitrarily if you have no criticism of doing it arbitrarily).
if the problem specifies that some solutions are better than others (not my preferred way to specify problems – i think it’s epistemologically misleading), then when you act you should pick one of the solutions in the highest tier you have a solution in, and reject the others. whether this method (pick a highest tier solution) is correct, and whether you’ve used it in this case, are both binary issues open to criticism.
when you guess it’s best to stop and your guess is non-refuted and the guess to continue looking is refuted. (you may, if you want to, define some stopping metric and make a subject-to-criticism binary yes-or-no judgement about whether to use that stopping metric.)
i think small kids do guesses and criticism, and use methods of learning (what I would call philosophical methods), even if they can’t state those methods in English. i also think ppl who have never studied philosophy use philosophy methods, which they picked up from their culture here and there, even if they don’t consciously understand themselves or know the names of the things they’re doing. and to the extent ppl learn, i think it’s guesses and criticism in some form, since that’s the only known method of learning (at a low level, it’s evolution – the only known solution to the problem of where the appearance of design comes from – saying it comes from “intelligence” is like attributing it to God or an intelligent designer – it doesn’t tell you how god/intelligence does it. my answer to that is, at a low level, evolution. layers of abstraction are built on top of that so it looks more varied at a higher level.).
It depends on what do you want to do with them. If all you want to do is keep them on a shelf and once in a while take them out, dust them, and admire them, then no, you don’t. On the other hand, if you want to persuade someone to change their mind, evidence might be useful. And if you want other people to take action based on your claims’, ahem, implications, evidence might even be necessary.
It seems that the root of these problems is your insistence that truth is a binary category. If you are forced to operate with single-bit values and have to convert every continuous function into a step one, well, sure you will have problems.
The thread seem to be losing shape, so let’s do a bit of a summary. As far as I can see, the core differences between us are:
You think truth (and arguments) are binary, I think both have continuous values;
You think intellectual debates are primary and empirical testing is secondary, I think the reverse;
Looks reasonable to you?
the two things you listed are ok with me. i’d add induction vs guesses-and-criticism/evolution to the list of disagreements.
do you think there’s a clear, decisive mistake in something i’m saying?
can you specify how you think induction works? as a fully defined, step-by-step process i can do today?
though what i’d prefer most is replies to the things i said in my previous message.
I would probably classify it as suboptimal. It’s not a “clear, decisive mistake” to see only black and white—but it limits you.
In the usual way: additional data points increase the probability of the hypothesis being correct, however their influence tends to rapidly decline to zero and they can’t lift the probability over the asymptote (which is usually less than 1). Induction doesn’t prove anything, but then in my system nothing proves anything.
What you said in the previous message is messy and doesn’t seem to be terribly impactful. Talking about how you can define a loss function or how you can convert scores to a yes/no metric is secondary and tertiary to the core disagreements we have.
the probability of which hypotheses being correct, how much? how do you differentiate between hypotheses which do not contradict any of the data?
For a given problem I would have a set of hypotheses under consideration. A new data point might kill some of them (in the Popperian fashion) or might spawn new ones. Those which survive—all of them—gain some probability. How much, it depends. No simple universal rule.
For which purpose and in which context? I might not need to differentiate them.
Occam’s razor is a common heuristic, though, of course, it is NOT a guide to whether a particular theory is correct or not.
Do all the non-contradicted-by-evidence ideas gain equal probability (so they are always tied and i don’t see the point of the “probabilities”), or differential probability?
EDIT: I’m guessing your answer is you start them with different amounts of probability. after that they gain different amounts accordingly (e.g. the one at 90% gains less from the same evidence than the one at 10%). but the ordering (by amount of probability) always stays the same as how it started, apart from when something is dropped to 0% by contradicting evidence. is that it? or do you have a way (which is part of induction, not critical argument?) to say “evidence X neither contradicts ideas Y nor Z, but fits Y better than Z”?
Different hypotheses (= models) can gain different amounts of probability. They can start with different amounts of probability, too, of course.
Of course. That’s basically how all statistics work.
Say, if I have two hypotheses that the true value of X is either 5 or 10, but I can only get noisy estimates, a measurement of 8.7 will add more probability to the “10” hypothesis than to the “5″ hypothesis.
what do you do about ideas which make identical predictions?
They get identical probabilities—if their prior probabilities were equal.
If (as is the general practice around these parts) you give a markedly bigger prior probability to simpler hypotheses, then you will strongly prefer the simpler idea. (Here “simpler” means something like “when turned into a completely explicit computer program, has shorter source code”. Of course your choice of language matters a bit, but unless you make wilfully perverse choices this will seldom be what decides which idea is simpler.)
In so far as the world turns out to be made of simply-behaving things with complex emergent behaviours, a preference for simplicity will favour ideas expressed in terms of those simply-behaving things (or perhaps other things essentially equivalent to them) and therefore more-explanatory ideas. (It is at least partly the fact that the world seems so far to be made of simply-behaving things with complex emergent behaviours that makes explanations so valuable.)
I don’t need to distinguish between them, then.
so you don’t deal with explanations, period?
I do, but more or less only to the extent that they will make potential different predictions. If two models are in principle incapable of making different predictions, I don’t see why should I care.
so e.g. you don’t care if trees exist or not? you think people should stop thinking in terms of trees and stick to empirical predictions only, dropping any kind of non-empricial modeling like the concept of a tree?
I don’t understand what this means.
The concept of a tree seems pretty empirical to me.
there are infinitely many theories which say trees don’t exist but make identical predictions to the standard view involving trees existing.
trees are not an observation, they are a conceptual interpretation. observations are things like the frequencies of photons at times and locations.
Isn’t it convenient that I don’t have to care about these infinitely many theories?
Since there is an infinity of them, I bet you can’t marshal critical arguments against ALL of them :-P
I think you’re getting confused between actual trees and the abstract concept of a tree.
I don’t think so. Human brains do not process sensory input in terms of ” frequencies of photons at times and locations”.
why not?
you can criticize categories, e.g. all ideas with feature X.
i don’t think so. you can’t observe entities. you have to interpret what entities there are (or not – as you advocated by saying only prediction matters)
Why not what?
How can you know that every single theory in that infinity has feature X? or belongs to the same category?
My nervous system makes perfectly good entities out of my sensory stream. Moreover, a rat’s nervous system also makes perfectly good entities out if its sensory stream regardless of the fact that the rat has never heard of epistemology and is not very philosophically literate.
Or not? Prediction matters, but entities are an awfully convenient way to make predictions.