From 5ae8300706c3aaf8bdfb6107a9034803f1db9324 Mon Sep 17 00:00:00 2001
From: Flancian <0@flancia.org>
Date: Wed, 14 Apr 2021 13:43:05 +0200
Subject: [PATCH] Auto: misc.

---
 notsof.md | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/notsof.md b/notsof.md
index d574e7deb..e83d12e2c 100644
--- a/notsof.md
+++ b/notsof.md
@@ -101,6 +101,11 @@
   - helps us define L and thus settle on the G(M, L)
 - three formal properties of G(M, L)
   - L describes all the interactions between variables there is. Because L is the set of two-variable correlations, M must be chosen so it's free from n-variable correlations with n > 2.
+    - superordinate concepts like "economics" and "acoustics" tend to be problematic here, and they should be seen as intermediate steps -- must be broken down
   - the two-variable correlations must be small for any pair.
   - p(xi = 0) should be the same for all i (?)
-   
\ No newline at end of file
+    - but in the next sentence Alexander moves to "roughly the same in significance", which is both distinct than the previous and more reasonable IMHO.
+   
+## Solution
+- the tension between form-making (integration) and analysis (fragmentation) is resolved by "finding knots": problems are not homogeneous, they actually have some structure, so there are right ways to tackle them.
+  - reminds me of [[computational complexity]]
\ No newline at end of file