virus: New Ideas

Dave Pape (
Thu, 22 May 1997 23:25:50 +0100 (BST)

People! Remember those crap posts I sent, one half-finished and the second
apologising for the first? Here's the third.

Soz if you've heard some of this before- if you have, the later bits are
newer. I just felt a couple of pennies dropping today, and fancied a memetic

Still talking about ideas and feedback and self-reference and self-replication.


Bunch of reactions in which one or more of the products catalyse/speed up
their own production. Looks like the more of this stuff you have, the more
you get. Proposed as a mechanism that preceeded DNA "self-replication".


If you've got a test tube full of DNA, suspended in pure water- no RNA, no
enzymes, no proteins, no phosphate- it won't make any more DNA. DNA doesn't
just "replicate /itself/". It's always needed loads of other chemicals to
help. There's a /system which replicates DNA/. The system's the cell, or
maybe the organism.

And I don't think cells self-replicate either. They need a wider ecological
system. This isn't me just being a raging holist, I'm trying to break down
the conceptual barrier between autocatalysis and self-replication. Cos I
think "self-replication" is actually a very intimate form of autocatalysis,
not a qualitatively different phenomenon.

Point is, there's this class of systems, AUTOCATALYTIC systems, in which
bits interact, and more copies of some of those bits get made. The systems
are self-referential, in that processes within the systems act on other bits
of the system. DNA's NOT a pure self-replicator, it's a component of a
neatly packaged autocatalytic system: a cell. RNA acts on proteins, enzymes
act on DNA, and more, and the rolling effect is more DNA.

And before cells, before DNA "replication", there was a gradual tendency
towards tightening up of self-reference/autocatalysis in the Earth's
chemical systems. EG growing nucleotide chains hanging onto clay crystals
and bathed in primeval soup- a proposed chemical mechanism that DNA
replication may have evolved from.

What we call self-replication is, IMO, very tight autocatalytic
self-reference. The cell is a serious fucking tightening of /chemical/


Went to a pub this week when it was Quiz Night. Seethe: the geezer doing the
quiz had a shite PA system, and his tinny, patronising,
loves-the-sound-of-his-own-voice voice rang a bit because there was slight
feedback in the speaker system.

"To the nearest 5 centimetres, what's the average yearly rainfall... in
Moscow?" ...wanker.

The feedback is his voice amplified by the speaker and his mic sending some
of that signal back to the speaker. The system takes part of its output as
part of its input. Refers to itself. Doug Hofstadter, we worship you.


Firstly, as far as I'm concerned a new idea emerges from interaction of
known ideas. The /known/ ideas are ones you'd use to explain or define the
/new/ idea. These known ideas are related, in that to learn a new concept
they have to activate together. So to teach memetics you hype up people's
known ideas about evolution, replication, Mind, culture, viruses etc, hoping
to get them interacting in the way they interact in YOUR head, and then get
your victim to label the emergent stuff... "memetics".

As the known ideas become associated together, activating one will activate
(some of) the others- and soon you get a feedback system set up cos idea A
activates B, which activates C, which activates A.

This feedback itself would, I think, have the effect of tightening itself
up. Because, neural-substrate-wise, ideas get associated together when
they're activated together, and associated ideas activate each other more.
The two tendencies tickle each other.

The more the interacting system of known ideas tightens up, and as mutual
activation gets more intimate, you get stronger emergent feedback, you get
it with weaker input to the system, and the feedback lasts longer before it

Like if Pubquizman stood closer to his loudspeaker, the ringing tone after
his voice would get louder, go on for longer, and he wouldn't have had to
talk so loud for the feedback to boot off, but his accent would still be

I think any new concept (meta-concept) IS such an emergent feedback signal.
Remember, we're talking crazy parallel neural systems here: the feedback's
not just an amorphous whiney noise, it'll be a complex neural activity
pattern, probably patterned in time as well as across cortex.

The hallmark of an idea you ALMOST know is that you fleetingly get glimpses
of understanding, which fade away dead quickly. Partial feedback which isn't
strong enough to be perpetuated in time for very long, because the input
wasn't loud enough, or the self-referential system it emerges from wasn't
tight enough. Like the fading ringing tones after Pubquizman's voice, right
about just before I KICKED his bleating hamster-like low-dignity-balding
head through the speaker grill, and stuffed the mic up his blinkered
suburban nose.

Which is when I found out that, if you put a mic right up against a speaker
it's wired into, it makes feedback that'll keep going by itself. The TONE
comes out the speaker, goes into the mic and back round the loop again.
That's what we call self-replication. The tone perpetuates itself- WITH THE
SYSTEM'S HELP (and remember, the signal's never utterly pure feedback).

...And that's the hallmark of an idea that you /know/: given a trigger
input, you get a cognitive activity pattern forming from very tight
self-reference in terms of ideas, and emerging from this you get what
appears to be self-replicating activity patterns.


When you join a team: to start with you need lots of help and advice, so
decisions about your job are effectively made by several minds. Over time
the interaction of job-related ideas tightens up, until eventually you
become, INTERNALLY, an autocatalytic system of "expert" ideas. Meaning,
you're an opinionated git, and seeing a situation provokes long streams of
thought about how to deal with it.


And more generally, I think bootstrapping a mature human consciousness
involves (not entirely, but importantly) generating an autocatalytic set of
very high-level ideas, which interact as a system with perceptual input to
produce the apparently self-replicating Self. Which isn't really a
self-replicator, but is the feedback bubbling out the top of the
memetic/conceptual interaction in your mind/brain.

Okay. I'm after comments. I know it's argument-by-analogy, but what the
fuck. It kind of deals with my former problems with ideas being defined in
terms of other ideas, so how can you think of AN IDEA? -I'd now say that the
criterion for saying "that's one idea" rather than "that's several ideas all
interacting" is this perpetuating feedback, maybe in terms of neural
activity patterns.

Whatever, please have a go cos it's all a bit vague and plastic still,
though the fact that I've been thinking about it all week and am sending it
to all my people must mean that I now "know" it- cos it's autocatalytic, see?

Dave Pape
I am ready.

Phonecalls: 0118 9583727 Phights: 20 Armadale Court
Westcote Road
Reading RG30 2DF