The Bongard issues on the 1 aspect of the Meta-Bongard are all simple to resolve (Numbers are OP’s reference system):
1A:
Nothing | Something
1B:
White form | Black form
1C:
Square | Non-Square Rectangle
1D:
Big factor | Small factor
1E:
Threesome | Foursome
1F:
Black Circle | White Triangle
The 2 aspect is considerably tougher… (Bold Numbers are Meta-Bongard field, Italic are Bongard)
2A:
Here, the 2 sides are precisely the identical, so there isn’t a attainable rule to distinguish them.
2B:
Here, there may be an obvious rule: Arranged | Scrambled.
Unfortunately, the order of the items can’t be used to differentiate the perimeters of a Bongard drawback, and the items themselves are in any other case equivalent.
2C:
This one is similar as 1C – besides that its 2D has been changed with a Square, breaking the rule and stopping a brand new rule from being attainable (because the shapes are in any other case equivalent).
2D:
Here, the 2 sides are virtually, however not fairly, equivalent – shades of 2A! There is an obvious rule right here, although: Bigger | Smaller.
Unfortunately, very like in 2B, this form of comparability is not a sound approach to make a rule, and the “small” issues are too shut in measurement to the “large” issues to only copy-paste the rule for 1D.
2E:
Here, either side are only a single dot, randomly positioned. There’s virtually a rule right here (Centered | Not Centered), however 2F breaks it.
2F:
Here, either side are simply noise. I’m positive a devoted laptop might tease a rule out of this one, nevertheless it will not be human readable and can really feel very contrived – greatest to not trouble.
So what is the rule for the Meta-Bongard?
I remorse to tell you that I cheated – I decoded @athin’s rot13 remark beneath the query. The rule is Valid | Invalid.
That is, the 1 aspect are all legitimate Bongard issues, and the 2 aspect are all invalid Bongard issues. I had an inkling of this, however 2C threw me for a loop (I could not see that it was completely different from 1C till it was too late).
