There is a word at the heart of memetics that looks almost innocent.
Replicator.
It sounds technical. It sounds precise. It sounds as though it names a kind of thing whose existence has simply been discovered.
But the word deserves more attention than that.
For replicator does not merely name an entity. It construes an entity through its relation to a process.
Something replicates.
We call it a replicator.
And once we have done so, it becomes remarkably easy to forget the process from which the noun came.
This is not an objection to the word.
Scientific language needs nouns. It needs ways of stabilising processes as objects of thought and discussion. Without such resources, it would be extraordinarily difficult to build cumulative explanations.
The interesting question is therefore not whether replicator is legitimate.
It is:
What kind of construal does the word make possible?
And once we ask that question, the apparently simple word opens onto a much larger problem concerning process, participant, agency and explanation.
Begin with the clause
Let us start as simply as possible:
The gene replicates.
At first sight, there seems to be nothing remarkable here.
The clause construes a material process in which the gene is a participant.
In the transitivity model, the gene can be analysed as Actor and replicates as the Process.
This is already enough to produce a familiar way of talking:
The gene replicates.
The gene appears to be the one doing the replicating.
But SFL gives us another way of looking at the same experiential configuration.
In the ergative model, the gene can be analysed as Medium.
That is an important difference.
The two analyses do not describe two different biological events. They provide different grammatical perspectives on the same experiential domain.
The transitivity analysis asks, in effect, about the participant that figures as Actor in the unfolding process.
The ergative analysis asks about the participant through which the process is construed as occurring, and whether there is an external Agent bringing the process about.
Consider the explicitly causative clause:
The cellular machinery replicates the gene.
Now the ergative analysis can distinguish:
Agent: the cellular machineryProcess: replicatesMedium: the gene
The gene has not ceased to be biologically important.
But its grammatical role has become clearer.
It is possible for the gene to be construed as Actor when the process is presented in an intransitive clause, while also being construed as Medium when the process is presented ergatively.
That observation should make us cautious about moving directly from Actor to agent in the philosophical sense.
The grammatical role does not establish an intentional agent.
Indeed, the ergative analysis makes visible something that an Actor-centred reading can easily conceal:
a participant through which a process occurs need not be the source of that process.
This distinction will matter.
The process becomes a thing
Now consider another way of construing the same domain:
The replication of the gene occurs through biological processes.
Something has changed.
In:
The gene replicates
the process is expressed congruently by the verb replicates.
In:
The replication of the gene occurs...
the process has been reconstructed as the noun replication.
This is the familiar SFL phenomenon of grammatical metaphor.
A process that was congruently realised by a verb has been reconstrued as a thing-like entity.
That transformation is extraordinarily important to scientific discourse.
Once we have replication, we can say:
replication is rapid;
replication is accurate;
replication depends upon cellular machinery;
replication produces variation under certain conditions.
The process has become something we can classify, compare, qualify and place into further relationships.
This is not a defect.
It is one of the great powers of scientific language.
Nominalisation allows discourse to accumulate experience into stable conceptual forms.
Processes can become objects of reflection.
Relations between processes can become propositions.
Complex explanations can be compressed into compact nominal groups.
Science could scarcely function without this capacity.
But there is a consequence.
Once a process has been reconstrued as a thing, it becomes possible to lose sight of its processual character.
Replication begins to look like something that exists in its own right.
And then we can ask:
What causes replication?
Where does replication occur?
What does replication produce?
What does replication require?
The nominalised process has become a participant in a new network of meanings.
This is often exactly what we want.
But it also means that the language can move us away from the unfolding event and toward a world populated by increasingly stable entities.
That is the power—and the subtle danger—of grammatical metaphor.
But replicator is not simply replication
Here we need to be careful.
It would be tempting to tell a simple story:
replicate → replication → replicator.
But that would collapse two different linguistic developments.
Replication is a nominalisation of the process replicate.
Replicator is a noun formed from replicate with the agentive suffix -or.
They do different conceptual work.
Replication gives us a thing-like construal of the process.
Replicator gives us a noun for an entity characterised in relation to the process.
That distinction is crucial.
We can therefore place the forms beside one another:
replicate — processreplication — process reconstrued as thingreplicator — entity construed through its relation to the process
The last formation is especially interesting because the suffix -or has an agentive tendency.
A generator is something that generates.
An operator is something that operates.
An accelerator is something that accelerates.
A replicator is something that replicates.
The morphological pattern invites us to construe the entity as the one responsible for the process.
But morphology does not by itself settle the ontology.
A generator need not be a conscious agent.
An accelerator need not possess intentions.
An operator in mathematics certainly need not be an agent in the philosophical sense.
The suffix supplies a particular way of construing the relationship between an entity and a process.
What we do with that construal is another matter.
Actor is not Agent
This is where our earlier distinction becomes essential.
In ordinary English, actor and agent can sound almost synonymous.
In SFL, they are not.
Actor belongs to the transitivity model.
Agent belongs to the ergative model.
And neither should simply be equated with philosophical agency.
This matters because the sentence:
The gene replicates
does not tell us that the gene possesses intention, consciousness or purpose.
It tells us how the process is being construed in the clause.
In the transitivity model:
the gene = Actor
In the ergative model, the same participant can be:
the gene = Medium
when the process is construed as one brought about through an Agent.
That gives us a useful warning against a common conceptual slide.
We can move from:
the gene is Actor
to:
the gene is a replicator
without noticing that we have changed the level of description.
And we can move from:
the gene is a replicator
to:
the gene is trying to replicate itself
without noticing that we have made a further interpretive move.
The grammar did not make that last move.
The word replicator did not logically entail it.
Something else has entered the discourse: a construal of evolutionary processes in terms of interests, strategies and persistence.
That is where the language of the selfish gene becomes especially powerful.
The agent hiding in the noun
The interesting feature of replicator, then, is not that it literally contains a philosophical theory of agency.
It doesn't.
Rather, it offers a linguistic resource that makes a particular construal easy.
The process is replication.
The entity associated with that process is the replicator.
The noun therefore invites us to ask:
What does the replicator do?
And the answer seems obvious:
It replicates.
The circle closes.
The entity is defined by the process.
The process is attributed to the entity.
The entity now appears to explain the process.
This is where we need to be careful.
A noun derived from a process does not necessarily explain the process.
Calling something a replicator does not yet tell us why replication occurs.
It tells us how we have chosen to classify the entity in relation to the process.
That distinction is easy to lose because ordinary language is extremely good at making classifications look explanatory.
If we say:
The gene is a replicator.
we may feel that we have identified an important mechanism.
And perhaps we have, depending on what exactly the term is intended to mean.
But the explanatory work lies in the account of the biological processes through which genetic material is copied, inherited, altered and differentially retained.
The noun itself does not perform that work.
Naming a pattern is not explaining it
This is a general problem in scientific reasoning.
Suppose we observe that something tends to occur repeatedly.
We give the pattern a name.
The name becomes convenient.
We then begin to talk about the named phenomenon as though it were a causal object.
Sometimes this is entirely legitimate.
Sometimes the named pattern corresponds to a real mechanism.
Sometimes it is a useful level of abstraction.
But sometimes the name merely packages the phenomenon we were trying to explain.
The danger is especially acute with nouns derived from processes.
Consider:
selection selects.
adaptation adapts.
organisation organises.
replication replicates.
The sentences are grammatically possible.
But they may tell us little more than the original process did.
A noun can give the appearance of explanatory depth without necessarily adding causal content.
Again, the point is not that nominalisation is bad.
It is that nominalisation changes what can be done with an experience in discourse.
It does not automatically change the experience itself.
The scientific virtue of reification
Yet we should resist an overly suspicious attitude toward this process.
Scientific knowledge requires forms of reification.
A physicist speaks of force, energy, field and momentum.
A biologist speaks of inheritance, selection, fitness and replication.
A linguist speaks of process, participant, clause, genre and meaning.
These terms allow complex relations to become stable objects of inquiry.
Without them, cumulative knowledge would be extraordinarily difficult.
The important distinction is therefore not:
reification good / reification bad.
It is:
What has been reified, and what happens to our understanding when we forget that it has been reified?
A nominalised process can become an extremely useful object of thought.
But the thing-like noun should not make us forget the process from which it was abstracted.
This is especially important when we cross levels of organisation.
A gene participates in replication.
A replicator is a way of characterising something in relation to replication.
But neither formulation by itself tells us the whole biological story.
The process extends through relations among genetic material, molecular machinery, cells, organisms and environments.
The noun is useful because it selects one aspect of that organisation.
It becomes misleading only when the selected aspect is mistaken for the whole.
The replicator and the disappearing system
This is perhaps the deepest problem.
The word replicator directs our attention toward a thing.
But replication is not a solitary event.
There must be a system in which replication can occur.
There are mechanisms.
There are substrates.
There are conditions.
There are constraints.
There are interactions.
There are consequences.
There is an environment.
There are processes of inheritance and differential persistence.
The replicator concept isolates one participant within this larger organisation.
That can be extraordinarily productive.
But it can also encourage a particular explanatory direction:
the replicator is there; therefore replication happens.
The system disappears behind the noun.
And once the system disappears, the participant can begin to look like the source.
This is exactly where the ergative perspective becomes philosophically useful.
The question becomes:
What happens if the participant we have made most salient is not the source of the process, but the Medium through which the process is realised?
The question does not deny the importance of the gene.
It relocates it.
The gene matters not because it possesses an independent power to reproduce itself, but because of the role it plays within a larger organisation of biological processes.
From genes to memes
Now we can see why the move from gene to meme is both ingenious and dangerous.
The analogy initially looks straightforward.
Genes are inherited.
Memes are inherited.
Genes vary.
Memes vary.
Genes persist differentially.
Memes persist differentially.
Genes are replicated.
Memes are replicated.
And therefore perhaps memes are cultural replicators.
But notice what has happened.
The word replicator has already done considerable conceptual work in the biological domain.
It has given us a way of selecting a particular participant in an evolutionary process.
When we carry the term across to culture, we carry much of that construal with it.
We begin looking for cultural entities that behave like replicators.
We expect them to be copied.
We expect them to compete.
We expect them to survive.
We expect them to spread.
And eventually we may begin to ask what the meme itself is trying to do.
The metaphorical extension therefore does not merely transfer a scientific term.
It transfers a way of organising the phenomenon.
This is where the distinction between biological and cultural inheritance becomes crucial.
What exactly is copied?
Suppose a person teaches another person a proverb.
What has been replicated?
The words?
Their sequence?
Their grammatical pattern?
Their sound?
Their meaning?
The circumstances in which the proverb is appropriate?
The cultural knowledge required to recognise when it is appropriate?
Or perhaps a potential for producing similar utterances in new circumstances?
The question becomes even more difficult with a concept.
If I explain evolution to you, have I replicated a piece of information?
Have I transferred a representation?
Have I reproduced a semantic pattern?
Have I enabled you to participate in a conceptual system in which the term evolution has a particular range of meanings and relations?
These possibilities are not equivalent.
And culture gives us a phenomenon that biological replication does not straightforwardly capture:
interpretation.
A cultural form can be received and understood differently.
It can be translated.
It can be adapted to a new situation.
It can be deliberately transformed.
It can be misunderstood in a productive way.
It can be embedded in a new practice.
It can acquire meanings that were unavailable in its original context.
The cultural form may therefore persist without being copied in any simple sense.
It can be re-realised.
That is a very different model of inheritance.
The replicator and the semiotic system
This brings us to the question we began to approach in our earlier discussions.
Perhaps the cultural analogue of the gene is not a cultural object at all.
Perhaps culture is better understood as a semiotic system.
Language is one such system.
Other semiotic systems operate alongside it and interact with it: visual forms, music, gesture, bodily practice, material design, diagrams, notation and technological forms, among others.
These systems provide organised possibilities for making meaning.
A child does not inherit a language by receiving a copy of English.
A student does not inherit mathematics by receiving a mathematical object.
A musician does not inherit a musical tradition by copying a physical token of it.
They enter practices in which the relevant semiotic resources are already organised.
They learn to participate.
And in participating, they reproduce the system while also changing it.
This gives us a very different conception of cultural inheritance.
The inherited thing is not necessarily a unit.
It may be a meaning potential.
The continuity is not necessarily the preservation of an object.
It may be the continued availability of a system of semiotic possibilities.
The transmission is not necessarily copying.
It may be participation.
Reproduction without replication
This suggests a distinction that may eventually prove central to our argument:
Cultural reproduction need not be replication.
A cultural form can be reproduced by being used.
A genre survives because people continue to produce texts recognisable as instances of it.
A ritual survives because people continue to enact it.
A language survives because people continue to use its resources.
A scientific tradition survives because people continue to participate in its practices.
A musical tradition survives because performers continually realise and transform its possibilities.
In each case, continuity depends upon repeated instantiation.
But the instances do not have to be identical.
Indeed, if every instantiation were identical, cultural systems would have no capacity for historical development.
What persists is not necessarily the token.
What persists is the capacity to generate recognisable new instances.
And that brings us closer to a semiotic account of inheritance.
The replicator was a useful question
Perhaps we can now see what was both right and wrong about the replicator.
It was right to insist that cultural change should not be thought of as a mysterious process unrelated to inheritance.
It was right to look for mechanisms through which forms persist and vary.
It was right to challenge the assumption that biological organisms are the only things whose histories can be described evolutionarily.
But the concept may have inherited too much from the grammar of its biological metaphor.
The replicator encourages us to look for a thing that reproduces.
The cultural world may instead require us to look for a system that makes reproduction possible.
That is a very different object of inquiry.
And it shifts the question from:
What is the cultural replicator?
to:
What makes cultural inheritance possible?
The latter question does not presuppose the answer.
It leaves open the possibility that what is inherited is a pattern, a practice, a semiotic resource, a conceptual relation, an institutional arrangement, or an organised potential for meaning.
It also leaves open the possibility that several of these participate together.
That seems closer to what culture actually looks like.
The word that started the trouble
The word replicator therefore turns out to be more interesting than it first appeared.
It is not simply a scientific label.
It sits at the intersection of several powerful linguistic resources.
A process can be construed congruently as a Process:
the gene replicates.
The same experiential domain can be reconstrued through grammatical metaphor:
the replication of the gene.
The entity associated with the process can then be named through an agentive derivation:
the replicator.
The replicator can subsequently be placed into a discourse of competition, persistence and interest:
the selfish replicator.
At each step, something has been gained.
The phenomenon becomes easier to talk about.
Relations become easier to stabilise.
Complex processes become easier to compress.
Patterns become easier to compare across domains.
But at each step there is also a possibility of losing sight of something.
The process can disappear behind the noun.
The system can disappear behind the participant.
The Medium can appear to be the source.
A grammatical role can be read as philosophical agency.
A metaphorical construal can begin to look like an ontological discovery.
None of these transitions is inevitable.
But each is available.
And that is why the vocabulary deserves scrutiny.
Beyond the replicator
We began with a question about cultural evolution.
Dawkins's answer encouraged us to look for a cultural replicator.
But perhaps the more we examine the word, the less confident we should become that cultural evolution requires such a unit.
The interesting thing about culture may not be that it contains things that replicate.
It may be that human beings participate in semiotic systems that enable forms of meaning to persist, vary and become available for further participation.
That would still be an evolutionary story.
There would still be inheritance.
There would still be variation.
There would still be differential persistence.
There would still be cumulative historical change.
But the explanatory centre would have shifted.
From the thing to the process.
From the process to the system.
From the isolated participant to its relations.
From replication to re-realisation.
And eventually, perhaps, from transmission to participation.
We should not rush there.
The replicator deserves one final concession.
It was a brilliant conceptual invention because it forced us to ask what an evolutionary process might require.
But a conceptual invention can outlive the assumptions that made it useful.
Once we have learned to ask the question, we are free to change the vocabulary.
Perhaps the most important thing that can be inherited from Dawkins is therefore not the meme.
It is the question of inheritance itself.
What persists?
What changes?
What makes continuity possible?
And what kind of system allows a form to survive precisely by being transformed?
Those questions take us beyond the replicator.
They also take us, unexpectedly, toward the symbol.
And there, the problem of cultural inheritance will become stranger still.
No comments:
Post a Comment