> At 08:28 06/02/97 -0800, Richard wrote:
>
> >Dave Pape wrote:
> >
> >>Because always having to think about which ideas you should propagate
> >>and
> >>discuss in order to benefit your social organisation puts an instant
> >>and
> >>heavyweight selection pressure on the memes which will flourish in that
> >>organisation's group memetic processing resource, meaning that the
> >>opportunities for novel DESCRIPTIVE ideas are curtailed for the purpose
> >>of
> >>promoting a lower bandwidth set of PRESCRIPTIVE ideas.
> >
> >That would be true if there were some finite number of possible memes.
>
> Well, presumably there's some (huge) number of possible ways for all the
> neurons in all the brains in the world to be arranged each quantum time
> unit, and presumably there are fewer of these ways which code for religious
> artistic endeavour rather than all possible artistic endeavour. So I
> actually think it IS true.
Interpretation gets interesting.
I'm reminded of a little paper I was exposed to in Intro to Expert Systems:
A neural-network computer [hardware] was being trained to autolearn
phonemes, and used light patterns as output. This was repeated a number
of times.
#1) This system approached zero measurable errors. [Finland 1, Japan 0]
#2) This system was completely immune to foreign accents. [Finland 2,
Japan 0]
#3) The meaning of the light patterns was totally different upon each
trial. [Finland 2, Japan @#?!]
=====
If we were to locate the TL15 Neuronic Synchronizer[tm], *you* would
*not* interpret my firing pattern of neurons the way *I* interpret it.
[Hint: We're on the border between TL7 and TL8]
//////////////////////////////////////////////////////////////////////////
/ Towards the conversion of data into information....
/
/ Kenneth Boyd
//////////////////////////////////////////////////////////////////////////