[FRIAM] anthropological observations

[uǝlƃ ☣]

Now Marcus is just being sadistic. >8^D

My own guess at a summary of Eric's stance is that where we see qualities, we can, at will, invert the vision and see quantities. Fontana is a great source for distinguishing construction from evolution. But for me, BC Smith [†] is better for maintaining an agnostic flippability (Necker cube) between objects vs fields, nodes vs edges, nouns vs verbs. The conflation Nick began with between expected value (an algorithmic reduction from a distribution to a singular thing/object) and a *quality* ... a qualitative feature of the world, parsed (registered in BC Smith's domain) from the ambience of the world is not inherently a bad thing. That conflation is not inherently bad because they're similar. They're both transformations from a field to a thing. The key is to be able to flip it back again, from a thing to a field.

It is our nature as pattern-recognizers to parse the ambience into things ... at least that *was* our nature before the modern math/physics *field* techniques began to seep into our intuition. Some of us who deal with fields/ambience/distributions all the time have begun to relax the harsh and immediate parsing. Engineers tend to simply be a bit lazy about it. The parsing happens, but they talk of approximations and epsilon as it goes to ∞ or 0. Mathematicians talk of duals, congruence, bisimulation, isomorphism, comutation, etc.

But I think it can all be adequately understood in terms of qualities vs quantities. Qualities like "wetness" are precisely the same as things like "frozen pond". Quantities like 32°F are precisely the same as processes like "if I walk on that, I'll slip and fall". The language each of us uses to grok this stuff is a choice. Eric provided a nice swath across several domains. Maybe too many. We're faced with the tyranny of choice. I'd treat it like a cafeteria. Pull the thread you understand best.

Bah! 316 words ... close enough, I say!

[†] On the Origin of Objects or, better yet Philosophy of Mental Representation ed Hugh Clapin.

On 4/18/20 9:25 AM, Marcus Daniels wrote:
> I think the point is that the relations or contingencies you mention can be cast as posterior probabilities from observed many-body correlations.  Distributional thinking works fine in that case too, it is just that some of those conditional probabilities get very close to 1.  Others relations are softer, only giving slightly favorable odds.   Still others can be modeled, having surprising <http://cds.cern.ch/record/154856/files/pre-27827.pdf> behavior.


-- 
☣ uǝlƃ