From: Douglas P. Wilson (dp-wilson@shaw.ca)
Date: Sat Mar 30 2002 - 18:23:53 MST
(A reply to earlier messages on memetic noise and silence machines --
somewhat pedantic near this top of the page, characterising kinds of noise
with some almost textbook definitions, and no humour to lighten the load,
but see near bottom of page for content more interesting to virus mailing
list people, noting extreme non-linearity of viruses).
On Thursday, March28, rhinoceros <rhinoceros@freemail.gr> wrote:
> Is it possible to define a concept we would call memetic noise,
> by analogy to the noise in a communication signal?
This is an interesting question and I think the answer is more than just
"yes, it is possible", since so many uninteresting or worthless things are
possible, but "yes, it is possible and potentially useful". But noise in
communication signals, a central topic in communications theory (or
information theory) is a deep topic, since there are many kind of noise.
Especially important distinctions are between analogue and digital noise,
where classical (Shannon and Weaver) communications theory is largely
analogue in focus, and between linear and non-linear communication channels,
which can be analogue or digital but when the linear/non-linear distinction
is made the context is mostly analogue -- though it is an important
distinction in a digital context too.
The corresponding distinction in kinds of noise is (roughly) between
additive noise and modulation, where additive noise may be subtracted or
cancelled out (see below on "silence machine" and cancelling out noise)
while modulation may be irreversible change.
> I would imagine it as a superposition of many insignificant
> meme complexes, and it would look like something having
> no pattern at all. What do you think?
The key word here is "superposition" which seems a straightforward analogy
to the use of the term from physics e.g. "superposition of waves", and we
should therefore look at the "Principle of Superposition" which applies to
systems or media where the superposition of signals (or waves, etc.) does no
t involve any interference between these signals (or waves), and note that
this is a textbook definition of LINEAR systems -- linear systems are those
that obey the Principle of Superposition.
The later message from "rhinoceros", sent on March 30th, and entitled
"Silence machine neutralizes noise" describes a machine for neutralizing or
cancelling out noise, which is only possible in a linear system or media,
such as air at relatively low (e.g. atmospheric) densities (and relatively
low, e.g. less than 120 decibels sound intensities, which amounts to the
same thing).
This implicit assumption of linearity is typical of much discussion on
similar topics and is indeed the fatal flaw in many such discussions --
linearity is perhaps the most important mathematical property I can think of
and such not just be implicit or assumed. See
www.SocialTechnology.Org/linearity.html for more information on linearity
and its relation to other concepts such as pseudo-invertability.
The most opposite concept to linearity is perhaps the "Butterfly Effect"
which suggests that a butterfly flapping its wings may lead to a hurricane
somewhat later, perhaps on the other side of the planet. This is nothing
but an extreme (though hypothetical) example of non-linearity, an effect
being out of all proportion to its cause. The proportionality of cause and
effect is another textbook definition of linearity that can be shown
mathematically to be precisely equivalent to the Principle of Superposition.
A far better example of even more extreme non-linearity is the spread of a
virus, especially a literal, non-metaphorical, virus, a strand of DNA or RNA
with a thin protein code. One single virus, precisely one, could and
probably has on many occasions managed to reproduce itself through the slave
labour of some conscripted cell, then by conscripting and destroying many
more cells managed to spread itself globally, usually being benign but
perhaps being like the AIDS virus and destroying millions of human beings,
being more destructive of human life than the worst hurricane.
The virus that is the nominal subject of this mailing list, an intentionally
created memetic organism, is and must be a creature of non-linearity,
intended to cause effects out of proportion to its causes, to reproduce it
self over and over again with the aid of human beings (including those not
exactly conscripted but who do not volunteer to help reproduce the virus but
do so without their knowledge, as people so often do with memetic
organisms).
Personally I retain a fondness for linearity which I relate to the desirable
property of transparency, an almost equivalent property, though a non-linear
system may transmit information correctly and thus seem transparent when
there is no other signal to interfere with the information in the first
signal -- when there is no noise.
I see some kind of linearity as an evolutionary choice made by social
animals, including human beings, which permits a fellow social animal, a
fellow person, to understand or visualise the causes of your actions from
observing the actions themselves, especially for linguistic actions -- that
is to say, it makes it possible for someone watching your actions or
listening to your speech to invert or pseudo-invert whatever processing you
did to the signal in transmitting it.
Quoting briefly here from my own web page at
www.SocialTechnology.Org/linearity.html : -- "I see linearity as an
evolutionary choice, which social organisms have taken -- the essence of
linearity is the ability to guess at the inputs given the outputs, even in
the presence of feedback (perhaps especially in the presence of feedback).
It is just the opposite to determinism, which is means the output is more or
less determined by the input. The ability to guess the input from the
output is -- at its best -- what I described on another [much older] page,
www.SocialTechnology.Org/narrate.html the ability to deconvolve the
distortions of personality, something that can only happen with linear
systems. So we have to be more or less linear to be able to function as
social organisms, and I suspect machines will have to become more and more
linear as well, with all the consequences of it".
Viruses, extremely non-linear, are not social animals and represent the
other evolutionary choice, specialisation for reproduction, not
communication. Social animals are those whose ancestors made the choice
of linearity, pseudo-invertibility, inference of cause from observation of
effect, especially inference or induction of linguistic inputs from
observation of linguistic outputs.
dpw
This archive was generated by hypermail 2.1.5 : Wed Sep 25 2002 - 13:28:45 MDT