Re: on imitation

singa (
Tue, 24 Nov 1998 01:57:00 +1100

Date: Tue, 24 Nov 1998 01:57:00 +1100
From: singa <>
Subject: Re: on imitation

> From: 1998 10:22:12 -0000

To mu suggestion that: "I would suggest that it [imitation] is a very
inadequate key".
"Paul Marsden" wrote:

PAUL: Pure imitation, or in its abstract sense replication does not
result in
> descent with modification (evolution) because there is no modification.
> Imitation (replication) is just one of the key processes in the evolutionary
> loop of replication, VARIATION and selection. The specific mechanisms in
> the social world of all these processes must be understood to account for
> the diversity and structure in the social world.

ALEX: Imitation = replication. This is a very useful clarification.
However, I would ask what it is in cultural terms that is being
replicated. Behaviour? If so I would still ask how we get the necessary
variety (and therefore change and evolution) that we can observe in
cultural systems? You have confirmed what I stated earlier, namely that
imitation/replication does not lead to evolution. So what does? If we
use imitation, how do we get variation and were exactly does 'selection'
fit into this scheme of things if our only 'choice' so to speak is

PAUL: Depends what you mean by uniformity - I see a whole lot of
diversity out
> there and that is what interests me. History is not just one damned thing
> after another, it's one damned thing INSTEAD of another. Selection is
> important.

ALEX: In cultural terms we have both diversity and uniformity - at two
different levels. Diversity occurs at the individual level where each
action (behaviour) is a very particular response (selection) to
particular conditions. The 'conditions' are the local environment and
the response is the repertoire of possible behaviours available to the
individual from interaction and exchange with other individuals.
Uniformity occurs at the collective level where there are clear
uniformities/similarities which can be noted between these many
individual responses. In any cultural system there are 'styles',
theories, customs, genres, paradigms, patterns of behaviour (or
meme-plexes) which describe the regularities which inform group
behaviour. Regularities which are the product of communication and
exchange between individuals within a particular cultural system
(science, music, art, literature, etc. etc. The cultural system is
defined by the medium of exchange involved). The function of memetics as
I understand it is to model the processes which take place in cultural

PAUL: ....there is a selection process based on power, who to imitate IS
a key
> question for memetics - IMHO.

ALEX: 'Power'? I'm afraid I detect here a slippage from the systemic (or
scientific) arguement to the social. I think there is a clash or
confusion of systems here. I don't think we can import arguments from
outside a domain to explain what happens within it. Either we describe
how a system self generates its own evolution and structure or we are in
an area of pure confusion.

PAUL: Nope, if you read Cloak, he suggests that only if there is no
> instruction (defined as the functional relationship between cue and
> behaviour) when exposed to a cue, a new instruction is likely to be
> replicated, i.e social learning. One task for memetics is to understand
> which cue's we are hardwired to attend to - this is where I think EP and
> memetics have something useful to say to each other.

ALEX: I have not read Cloak, but it seems a somewhat atomistic (and
therefore restricted) view of cultural activity which if it means
anything means the availability of a collective repertoire of acts which
can be selected and re-combined to suit particular circumstances. How
else does language work for instance? We don't 'replicate' or invent new
'instructions' every time we meet a new event. How uneconomical. By the
way, 'cues' cannot be hardwired into anything. They are, logically,
signals from 'outside' - from the environment to which an individual
responds. Inevitably they are unpredictable in character. This is what
happens when we de-contextualize the system. In cultural terms there is
no such thing as the isolated individual. Culture and therefore memetics
(though its hard to believe) defines an ecological system-environment
relation with a communicational relation between them. What can 'social
learning' mean if not exchange (of experience) between many ndividuals
(the group)?

PAUL: No, the whole point about culture is that there are many diverse
> repertoires out there, which when they interact produce new relationships
> between cue and behaviour.

ALEX: Yes. As I have also said, 'there are many diverse cultural
repertoires out there' (architecture, politics, philosophy, sexual
relations, and so on), but they do not interact in the way you think
they interact. They each have different languages (media of expression).
How, for instance does music affect your science? I cannot build
buildings with words, only concrete and steel. As I suggested above 'cue
and behaviour' is too Pavlovian and mechanistic an approach to handle a
communicational system as complex and context-dependent as culture. You
cannot get evolution (an elaboration of a program) out of that. You can
only get 'more of the same' change.

Paul: No, imitation (replication) is a necessary but not sufficient
condition for
> evolution.

Alex: Yes. But,culturally the imitation of what? In fact there is a way
out of this somewhat literal (literary?) trap. By assuming that what is
imitated is NOT the specific form of behaviour - the purely analog or
chimpanzee mode as I would call it (which gets us nowhere), but rather
an event-in-context - a whole set of relations. This problem can be
easily solved if we get rid of the loaded term 'imitation' (or
replication) and substitute the idea (and process) of modelling. In this
the issue is one of translating a set of relations in one 'langauge'
into a formal system (another language) which we can manipulate. The
'cue', if you still want this idea, is built into the context which is
inherent to the concept of modelling. It is no longer an isolatable
entity or single signal, but a complex of conditions emanating from an
environment. (This is observable in everyday life. A single word or
gesture cannot be a cue. It might mean the opposite of what you think it
means. It has to be read as part of a context or collection of signals
or signs.

PAUL: Don't mistake a failure of imagination for an insight into
> Perhaps "There is no way" but we are chipping away at the awe-inspiring
> edifice of culture, and one day I think we will be able to explain some of
> it using the evolutionary loop of replication, variation and selection. I
> hope that my work on imitative suicides and imitative sociopathic behaviour
> from an evolutionary perspective might allow for insights understanding,
> explanation and prediction, not afforded by other perspectives.

ALEX: Yes, but that edifice has already been scaled from so many
different directions. There are so many available and useful theories
which describe the processes of cultural evolution which can be
integrated into a truely coherent theory that this 'chipping away' has a
somewhat nostalgic ring to it. Been there, done that. I think part of
the problem is that memetics has been from its origins trapped in a
so-called 'scientific' (in fact: atomistic/technological and mechanistic
mode derived from authority figures like Dawkins, Dennett et al, who
themselves do not take these literalist positions). This is a rather
dated position considering that what we are dealing with in memetics is
information systems, logical levels, complex symbolic systems, orders of
learning, ecological relationships and the analysis of context-dependent
systems. We can with some effort and more economically explain culture
through evolutionary processes, but not by ignoring the fundamentally
collective/communicationaland symbolic nature of the animal.

PAUL: I'm not sure I know what you mean, but if you are saying that the
> symbolic/representational nature of culture precludes the operationalisation
> of an evolutionary model of culture then I do not agree. Whilst I don't
> agree that this is the most useful way forward for all the reasons Derek and
> I have been debating with Aaron and Mario some big heavyweights in ECT
> (evolutionary culture theory) do agree with Aaron and Mario namely Durham,
> Sperber and Rindos.

ALEX: NO. I'm not saying that. Quite the opposite. The
symbolic/representational processes involved in cultural formation are
quite amenable to systematic enquiry (eg. modelling,
selection-combination processes, communication and exchange,
analog/digital differentiation, the idea of the message-in-circuit
(Bateson), meta-system transitions and so on). However they cannot be
understood using the techno-science appraoches currently used.

PAUL: They [cultural codes] might have non-linear emergent properties
but transmission is a
> discreet event.

ALEX: Yes. Of course.

PAUL: Okay so what do you want to do, study systems full of people
> radiating emergent properties and stop there? What we need is emergence
> without magic as Rob Clewley's excellent paper suggests. I'm a reductionist
> because I don't want to pin theories of culture on sky hooks, I want cranes
> to build with. Science is methodologically reductionist because it is a
> method, it cannot be anything else but methodologically reductionist, but
> that does not necessarily imply ontological reductionism.

ALEX: Sky hooks? Cranes? Where have I heard this before? AS I was saying
about authority figures.....! Anyway - 'emergent properties' are not
entities separate from the system from which they were derived. This is
a mistake carried over from the old Vitalist position and in part due to
the limitations of the English language. In an operational sense,
emergence refers simply to the reorganization or coordination of an
antecedent system. Nothing is added. The meta-system is exactly this. It
is the establishment of a new level (the meta-level) of controls or
rules derived from an assimilation of previous behaviours. IT is not
magical. It is systematic.

PAUL: Explanation requires reduction,otherwise you leave out what needs
to be
> explained, and the reductio ad absurdum counter to this does not work
> because our domain assumptions are the causal logic of
> self-emplacement/replication.

ALEX: 'Explanation requires reduction'. Of course, since we are faced
with the infinite complexity of the natural and cultural worlds.
However, the limits to reductionism are determined by the pragmatics of
explanation. Namely what level of the system am I trying to explain. It
is entirely a question of relevance. I am sure that the atomic structure
of the table on which I have my computer is of significance to some
explanation or other, but not to me using my computer. (The story goes:
American and North Vietnamese generals meet after the war to exchange
experiences. The American generals say: "we beat you guys in every major
set piece battle of the war". The North Vietnamese generals replied:
"Yes. This is true, but it is irrelevant"). The model must be adequate
to the complexity of the system to be modelled.

> One might
> > think therefore that there would be a shift upwards in scale to the
> > communicational systems level to

PAUL: This sounds like an interesting memetic project.[AB:....look at
the regularities which emerge
> > from the production and exchange of many messages - such as the emergence
> > of meta-systems. Or, perhaps a more elaborate theory of the constraints
> > imposed by rule-governed behaviour? The digital communicational processes
> > of selection and combination? Complexity theory, self-organization/autopoesis?]

ALEX: It is the memetic project.

PAUL: Try evolutionary World Systems Theory or Sanderson's Evolutionary
> Materialism - both have a stab at this.

ALEX: Not quite. These are essentially historical endeavours. K-waves
and things. See World Systems web site and Philofhi web sites.

PAUL: Yours is probably bigger than mine, but does that preclude me
from getting
> on and doing my own smaller micro-theoretical activity? Culture may be big,
> but it only exists as a product of individual acts (unless you are not a
> physicalist of course... and physicalism does not preclude emergent
> properties), and useful insights can be made from one on one activity. Size
> does matter, but it's what you do with it that counts. Of course yours
> might be producing vast quantities of emergent properties, and i may be
> shooting blanks - but at least I hope to be able to explain why....
> Paul Marsden
> Graduate Research Centre in the Social Sciences
> University of Sussex
> e-mail
> tel/fax (44) (0) 117 974 1279

ALEX: Paul says:"Culture may be big, but it only exists as a product of
individual acts..." Absolutely, but this 'product' is a product of
communication and exchange between the individuals within the system. If
we don't study this - the process which generates the (collective)
cultural codes we SAY we are looking at and the fundamental difference
between the analog communicational processes of animal behaviour and the
digital selection - combination processes of human systems, we will
still be chipping away at this edifice for another thousand years.
Memetics may be seen in part as the search for the discrete units which
are combined and recombined to make up the regularities of these
cultural codes. AS always, language is the perfect analogy.

Regards (and thanks to Paul for his comments)

Alex Brown

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)