Re: Memes and associative learning in neurons

Tue, 3 Nov 1998 17:46:35 -0500 (EST)

Subject: Re: Memes and associative learning in neurons
Date: Tue, 3 Nov 1998 17:46:35 -0500 (EST)

On Tue, 03 Nov 1998 23:38:15 +0800 Steve <> wrote:

> I think that one of the more peculiar questions that
> periodically pops up, that I've seen raised by others from similar schools
> of thought is "what is the purpose of consciousness?" I'm never too sure
> what this question, in the *sense* that it is asked, implies - are they
> wondering if the "purpose" of consciousness is to support the selfish gene?
> But then, what is the purpsoe of the selfish gene? Seriously... Like those
> who believe that God created the world. Who, pray tell, created God?

'Selfish' genes (I don't much like the term, but it's so standard
we'll stick with it) don't need a purpose, once in existence they
self-perpetuate. So I suppose you might say their purpose is to
survive and reproduce. Consciousness does need a purpose, because it
probably has a high cost in terms of energy etc., and in any case
unlike genes consciousness isn't a replicator. Its purpose is, I would
say, to maintain those selfish genes that constitute the human genome.

William James had a lot of interesting things to say on this subject.

I said:

> Brains associate
> experiences, but neurons don't (certainly not individually...).

and Steve replied:

> And *you're* going to need some evidence for this one.

No, the onus is on you, because your theory is the unorthodox one. We
know that neurons have some electrical properties. We don't know at
all that they can 'associate' things. That is frankly an outrageous
suggestion, and I don't have to disprove it. You have to prove it, or
at least provide some suggestive supporting evidence. (I'll accept any
supportive _evidence_ that you can offer, even if circumstantial).

> Fact....
> single-celled organisms (and that includes neurons) are alive.


> What
> does it
> mean to be "alive"?

Two orthodox suggestions:

Replicating information? - viruses are therefore alive, and perhaps
computer viruses too if you stretch it....

Metabolic activity? - viruses are not alive, only cellular organisms
are alive.

> What are some axioms we might apply to living entities?
> I'll provide one (playing devil's advocate)... all living entities make
> choices from their ecologies

That's not an axiom. That's a profoundly subversive theory which
requires justification.

> (I anticipate that you'll instantly disagree
> here, because you'll probably be basing your assumptions in some sort of
> biological programming/genetic determinism).

Dead right. I think you've got me to a tee.

> The fact is, that
> single-celled organisms are alive. A neuron is a single-celled organism.
> When you assert that neurons don't associate experiences, you're simply
> parroting a dogma that has no foundation in fact.

Err... well no, because it seems highly unlikely that a single neuron
is complex enough to do something like associate. Brains (even neural
nets in silico??) may associate, but single neurons? What you are
suggesting is a bit.., well homuncular really, is the word that springs
to mind. To say that individual neurons do something that only brains
are known to do is rather like saying that there are miniature brains
within the big brain.....

As for parroting dogma, that's what the dear British taxpayer pays me
for. But seriously, on a cellular level, there is no evidence of
quasi-cognitive properties in individual neurons. It's dogma because
it's the best current guess....

> Ok, let's say you argued the opposite veiw that all neurons 'need' to be
> able to do is fire at the approrpriate time and at the appropriate level.


> I'll argue that it is impossible for an assemblage of neurons to function
> in this manner. The reason being that it is too improbable to be able to
> sustain constructive dynamics in a system where participant entities
> (genetically programmed, no less!) are not "motivated" in some way....
> well, to be more specific, plainly ridiculous.

Try reading Daniel Dennett's book 'Darwin's Dangerous Idea'. I think
he answers your argument better then I ever could. What you're
proposing is the sort of concept that Dennett calls a 'shyhook'.

> For starters, how do
> they
> *get* to the right configuration (and please don't tell me that it's all
> programmed in the genes)?

Okay, I won't say _all_ in the genes, but a lot of it. Try 'Neural
Darwinism' by Gerald Edelman. He has some fair ideas about how brains
can develop without total genetic blueprinting.

> Look at the complexity of a human city, that
> has
> all the analogous characteristics of a living entity.

Hmmmm..... not too sure about that.

> Such complexity
> is
> possible only because the human participants constituting the Great Beast
> are players in a network of rewards and punishments. They are *motivated*
> in some way, to play the game. If a city's people werne't motivated in the
> right, culturally appropriate manner, there would be no city.

Now this is the correct subject matter for memetics. If you were to
confine yourself to socio-cultural phenomena, then we'd have no

> But imitation of concepts leads to the replication of concepts. Imitation
> is a subset of conceptualisation.

Firstly, I don't think that concepts are imitated (contentious issue,
plenty of people on this list will argue against me on that), but even
if they are it doesn't follow that imitation is a subset of
conceptualisation. Concepts might be just one of the several things
that are imitated.

> I wasn't aware that phenomenology was unfashionable.

> If it is, then so
> what? Just because some people disagree with phenomenology does not
> constitute proof that it is wrong.

Well, frankly the problem that most scientists have with Husserl is
a) he was fairly anti-scientific in his attitudes, I mean he was rather
disapproving of 'science as a means of knowing'.
b) it's difficult to see how you can scientifically approach a
subjective state. You might be able to have a go, but it will be
difficult to be quantitative - and that's absolutely esential if you
are going to have any evolutionary science.

> Heidegger's ideas are certainly
> relevant
> in my reasoning, and his philosophy is alive and well and doing the rounds.
> If it is not relevant to the view of life you work with, then perhaps we
> should simply agree to disagree.

Well, okay then, but if you can square Heidegger with modern science,
then you'll have performed an intellectual feat which will make you
famous. I'm serious, honestly, if you can do this it will
revolutionise science. Of course I also have to say I think it is
impossible, but I can't help wanting to wish you luck.

> The world of genetic determinism, one that assumes the brain to be some
> sort of biological computer, is a highly improbable one.

Well, actually I am very suspicious of the computer analogy of the
brain, while still being a genetic determinist of sorts. But I

> It is
> analogous to
> our current notions of the big bang (multiple universes, etc) and born of
> the same lifeless beast. The multiple universe version explains away the
> awesome improbability of life through the anthropic principle. That is,
> sure life is improbable, but we (the awesome improbability that we are) are
> here to witness our inevitability. But you see, I'm tackling all this from
> a completely different angle, with my basis being that life is inevitable.

Hmmmm.... Try 'Wonderful Life' by SJ Gould. If that doesn't convince
you that life is totally contingent, then nothing will.

> And no amount of hand-waving asserting that it's all programmed in the
> genes is going to convince me otherwise.

Hand-waving assertions are bad, I admit, and I hope I'm not doing them.
I just think that your theory needs squaring with accepted views on how
neurons work.

I said:

> >
> >But were and what are these neurons? This is 'grandmother neuron'
> >territory.

and Steve replied:

> Say what? I don't understand your question. They're appropriately located
> in the brain (!?) Neurons (in the visual cortex, for example) have
> their
> specialisations, one of which is related to length, another to
> darkness/brightness (or some other simple standards), etc. Where's the
> difficulty?

You'll have to be a bit more specific. Exactly where in the cortex are
these neurons that are 'related to length'? I don't think there are

> It is difficult, if not impossible, to *prove* that principles of
> associative conditioning apply to neurons to enable them to subjectively
> "interpret" their warm, wet worlds (and just as impossible to *disprove*
> it, I might add).

Try the following substitution:

> It is difficult, if not impossible, to *prove* [astrology] (and just
> as impossible to *disprove*
> it, I might add).

This is a poor line of argument. It's one that people proposing wild
theories (please don't feel insulted, but your theory is actually
quite wild, honest I'm not getting at you personally) often use.


This was distributed via the memetics list associated with the
Journal of Memetics - Evolutionary Models of Information Transmission
For information about the journal and the list (e.g. unsubscribing)