Corpus_Collosum leads us up from
neurons to "
complex processing systems" to
networks to
language centers to
cerebral hemispheres to
people,
teams,
organizations to
alliances to
political parties,
governments and all the way to
states and the
UN.
Most of the steps I view as requiring serious empirical proof (a scientific process which involves more than pointing at something I don't understand and telling me "emergent behaviour"). But I'll tackle that on some other battlefie^Wnode. Here I'm not so much after "intelligence" as after "fractal".
None of the above hierarchy is a fractal, of course! Fractals are exceedingly well-defined, and some qualititative statements of their properties have almost reached the status of definitions themselves. So let's take the best known characteristic: self similarity at all scales. If I magnify and show you a small enough piece of the Sierpinski triangle or Koch snowflake fractals, you cannot tell the magnification! (If I magnify and show you a small enough piece of a Julia set or the Mandelbrot set, then while it is possible to "find" it uniquely within the set if the picture is exact, it's not possible to do this if there is any potential error in the presentation, no matter how small)
Is this what Corpus_Collosum is claiming about his hierarchy? NO!. He's claiming the exact opposite!
If I look at a neuron, I don't see the UN or anything like it. If I look at people, they don't look like neurons (or behave like them). If I look at a company, it's nothing like its constituent employees.
That's the whole point of talking about "emergent complexity"! And it's the exact opposite of talking about fractals.
So, while fractals are definitely trendy and cool (and chaos is sometimes even relevant to some work on neural networks), mixing them with emergent complexity is, shall we say, unfounded.