I must warn you that this (in)famous problem of induction (the puzzle of grue) is rather confusing, and will likely take some time to get it. But as a system of induction/observations, it's striking how it matches with exactly how Socionics goes about creating knowledge.
Here are some videos on explaining the problem:
(From 3:15)
Basically, translated into Socionics, it goes like this:
Before observation (T), All observations of Fi is about relationships
OR
After observation (T), All observations of Fi is about getting the common cold easily
(T) = any arbitrary date set in the future
Before (T), All observations of Fi is about relationships
Therefore, Fi is about relationships
= Makes perfect sense!
After (T), All observations of Fi has been about getting cold easily
Therefore, Fi is about getting cold easily
= So is Fi about relationships, or is it about getting cold easily? Contradiction!
Arguments against this problem:
"But it's OBVIOUS that Fi being about relationship is TRUE, while getting cold easily is FALSE!"
- But how do we know that? We don't know that. It's only "obvious" since it's a common sense problem for us. There's no "justification" for believing that, since this is all based on observations. And no amount of observations can justify that belief.
"NO NO NO, I don't believe it! Fi is about relationships, and it can't be anything else!"
- Then you are saying that you can look into the future, somehow, even though you are looking into the past for patterns.
"Well, how about Fi is BOTH about relationships, AND getting cold easily?"
- This seems to create a problem of identity. Why should Fi contain things that have no connections with each other? What would be the purpose of Fi then?
"Well it may be so, but I don't really care either way... I'm just going to keep using Socionics anyway"
- This cannot be helped... but this problem can't simply be ignored or tossed aside, if we wish to progress.
Conclusion:
- The conclusion is perhaps, which can we say is the LAWLIKE generalization? Is it about relationships, or is it about getting cold easily? We won't ever know from induction.
This is why induction doesn't work, and rather deduction is required, on explaining *how* Fi works to FIND a causality, which IS lawlike. Everything else is more likely than not based on false pattern-recognition, confirmation bias, etc. We'd need a bottom-up approach, so to speak, of explanations, instead of a top-down approach of observations.
Here are some videos on explaining the problem:
(From 3:15)
Basically, translated into Socionics, it goes like this:
Before observation (T), All observations of Fi is about relationships
OR
After observation (T), All observations of Fi is about getting the common cold easily
(T) = any arbitrary date set in the future
Before (T), All observations of Fi is about relationships
Therefore, Fi is about relationships
= Makes perfect sense!
After (T), All observations of Fi has been about getting cold easily
Therefore, Fi is about getting cold easily
= So is Fi about relationships, or is it about getting cold easily? Contradiction!
Arguments against this problem:
"But it's OBVIOUS that Fi being about relationship is TRUE, while getting cold easily is FALSE!"
- But how do we know that? We don't know that. It's only "obvious" since it's a common sense problem for us. There's no "justification" for believing that, since this is all based on observations. And no amount of observations can justify that belief.
"NO NO NO, I don't believe it! Fi is about relationships, and it can't be anything else!"
- Then you are saying that you can look into the future, somehow, even though you are looking into the past for patterns.
"Well, how about Fi is BOTH about relationships, AND getting cold easily?"
- This seems to create a problem of identity. Why should Fi contain things that have no connections with each other? What would be the purpose of Fi then?
"Well it may be so, but I don't really care either way... I'm just going to keep using Socionics anyway"
- This cannot be helped... but this problem can't simply be ignored or tossed aside, if we wish to progress.
Conclusion:
- The conclusion is perhaps, which can we say is the LAWLIKE generalization? Is it about relationships, or is it about getting cold easily? We won't ever know from induction.
This is why induction doesn't work, and rather deduction is required, on explaining *how* Fi works to FIND a causality, which IS lawlike. Everything else is more likely than not based on false pattern-recognition, confirmation bias, etc. We'd need a bottom-up approach, so to speak, of explanations, instead of a top-down approach of observations.