The concept of logic meshes

[Betonhaus]

I'm wondering if this is a concept that anyone else is familiar with, and if they know a more proper name for it? Logic meshes as I understand it is the idea that ideas and concepts aren't considered true because a specific idea/concept supports it and so on like an infinite stack of turtles, but is considered true because it is connects to multiple concepts and ideas and they cosupport each other, forming a mesh of interconnected ideas and concepts that act as a filter that newly introduced ideas are filtered through and are either accepted or rejected by the mesh based on if it supports and is supported by the concepts in the mesh.

That doesn't meant that any individuals internalized logic mesh is the absolute truth, as it is trained and distorted by our education growing up and the media we consume. If we're lucky we learn how to form a robust logic mesh that can accurately filter out misinformation, but given enough time and a persistent source it's possible to accumulate enough misinformation that someone's logic mesh will actively reject truthful and accurate information without even attempting to analyze it. Some concepts can become so warped that information that refutes it would require reorganizing a significant portion of the logic mesh to accept it.

 

[Uncle Sid]

You might find this interesting: https://www.goodreads.com/en/book/show/287528

Although give the reviews a read - they somewhat express what I think you're getting at - that at higher levels, logic becomes self-referential instead of absolute.

 

[Milkis]

Sounds kind of like the concept of explanatory and predictive power in hypotheses. Theories are valuable because they let us make predictions about the world that bear true.

 

[Uncle Sid]

Theories are valuable because they let us make predictions about the world that bear true.

"All models are wrong, but some are useful" - George Box.

 

[Strix454]

I'm wondering if this is a concept that anyone else is familiar with, and if they know a more proper name for it? Logic meshes as I understand it is the idea that ideas and concepts aren't considered true because a specific idea/concept supports it and so on like an infinite stack of turtles, but is considered true because it is connects to multiple concepts and ideas and they cosupport each other, forming a mesh of interconnected ideas and concepts that act as a filter that newly introduced ideas are filtered through and are either accepted or rejected by the mesh based on if it supports and is supported by the concepts in the mesh.

Its not really a "logic mesh". It sounds like applying certain ideas out of machine learning to humans. Inference rules are applied to a "knowledge base" to generate a result. In such a system, what you call "the mesh" draws conclusions based on the sum of its learned knowledge. It doesn't really "filter" anything. It doesn't know "truth" either. It simply can draw conclusions based on the sum of its total learned knowledge. Is the sum of my knowledge consistent with the statement rather than asking if the statement is true or false.

Trying to apply models out of formal logic and machine learning/AI to humans & human thought has always been extremely problematic. But people are always doing it anyway.

At the root of some of it are a series of fallacies around education, "news" and the idea that there is "one right answer" to every question. That if a group of humans were all given the same information and same education, they would all come to the same conclusions about everything. That logic and reason can make democracy, politics and more generally most of human activity unnecessary.

 

[Betonhaus]

Its not really a "logic mesh". It sounds like applying certain ideas out of machine learning to humans. Inference rules are applied to a "knowledge base" to generate a result. In such a system, what you call "the mesh" draws conclusions based on the sum of its learned knowledge. It doesn't really "filter" anything. It doesn't know "truth" either. It simply can draw conclusions based on the sum of its total learned knowledge. Is the sum of my knowledge consistent with the statement rather than asking if the statement is true or false.

Trying to apply models out of formal logic and machine learning/AI to humans & human thought has always been extremely problematic. But people are always doing it anyway.

At the root of some of it are a series of fallacies around education, "news" and the idea that there is "one right answer" to every question. That if a group of humans were all given the same information and same education, they would all come to the same conclusions about everything. That logic and reason can make democracy, politics and more generally most of human activity unnecessary.

A machine learning system never really forgets or rejects incompatible statements. The human mind has already been demonstrated to regularly prune information, and has difficulty retaining information that doesn't relate to existing information (hence the need for mnemonics and such). But yeah machine learning does have a soul-searchingly high level of similarity to how the human brain functions.