2003-07-03

Parts, distance, relativity


p73 of Against Method

the rock stays with the tower, why?

or rather, why is our folk physics wrong?

something to do with the way we divide things into *parts*, that's really
tricky

so we solve it gradually by assuming a kind of inheritance, or rather the
system of parts we have mean that we have to assume a kind of inheritance in
our worldview for everything to start making sense


um


maybe it's this


because we divide the world into parts there is the *appearance* of qualities
of bits of the world leaking into each other; we're initially mislead and we
solve the problem by some form of relativity (the rock inherits the movement of
the tower because they are both moving round the earth), but that itself leads
to problems, etc

all because our chunking into parts - critical part of *naming*, semantic
handles - causes problems


(mostly i have respect for folk psychology, folk physics, etc. they're pretty
good. actually, i need a new word here. 'folk' is good for understandings of
the way the universe works, but i need something closer to how things are
conceptualised, or rather how the folk X is used in implicature, when the
meaning of X doubles back and helps the brain make an assumption: so distance
meaning causality evaporation is one of these. something to do with parts is
another.)


anyway, we chunk badly, then at some level of complexity introducing
relativity/inheritence because important: this happens *locally* (that is, over
small distance)


(it's important to remember here that distance is measured by an ordered list
of difference. the more steps of difference things are, the more distance is
between them. but the things themselves can't be compared: just whether they
are ranked before or after each other)


in software architecture:

code was fine for a while, then lots of it came around and we observed that
things would *move* together (there was not much distance between them,
transformations would propagate. think of the snowflakes again: how would they
be clustered? you can't cluster them as the edges of the clusters are too weak,
but occassionally slabs of snow *move* together. like that).

so we needed to introduce relatively. or rather, build it in to help code close
together to move together, so things wouldn't break.

this is object orientation: a misconception of relativity that says that things
inherit.


i wonder how often this problem with virtual worlds/ this pattern has been
experienced? and what is its analogue in cyberspace (it's not the addressing
scheme, that's not the distance involved. think about how the parts are
chunked.)