Mario Vaneechoutte (
Mon, 10 Feb 1997 09:25:38 -0800

Message-Id: <>
Date: Mon, 10 Feb 1997 09:25:38 -0800
From: Mario Vaneechoutte <>
Subject: Replicators

The posting of Nick Rose was a very interesting one because it
illustrates nicely some of the existing confusion: replicators,
Lamarckism, memotypes, .... It must be possible to get out of this
confusion with the public of readers we have here on this list. Many
items of evolutionary genetics are obviously ill understood (I don't
claim I understand them all), and I think that the usage of words like
replicator or other fancy words causes more confusion and problems than
it solves. Before we can find out which models from genetic evolution
could be used to understand cultural phenomena, we should try to get out
of this confusion.

Let me first say that I can very much agree with Dawkins that genes are
the unit of selection. But what about calling genes 'replicators'? What
about the phenotype-genotype distinction in biology? What about
'vehicles' and 'interactors'? Are these concepts really useful?

I think that the use of 'replicator' for a gene (I understand for a gene
the material nucleotide sequence, which is the encoding of genetic
information) is ill-chosen. This is also true in case you use 'gene' to
refer to the genetic information. Maybe we should also have a consensus
on the use of such a basic term: I'd prefer to use gene for the material
encoding and genetic information for its informational content.
Physical genes contain information about how replication should proceed,
but when you use '-ator', you refer to an active agent, and genes are
only a material instantiation of information on how to construct such an
agent. Genes do not replicate, they are replicated.

Also in case you talk of genes as the genetic information: can
information replicate? The only thing I see is that more material
instantiations of some specific information can be made in case the
resulting processes which it stands for appear to be successful in a
given selective environment.

There is more to say about the gene as a replicator, much more.

We always claim that genes are inherited, but a bunch of genes not
encapsulated by the cellular machinery wouldn't replicate very much.
Indeed, all organisms start as a single complete cell, and not as a
stretch of genes. This cell contains a genetic code but it also contains
enzymes needed to translate the code and to make a duplicate of it. So,
we need to inherit enzymes as well. Certainly eukaryotes need to inherit
all kinds of organelles. And mostly overlooked is the membrane which has
to be inherited. Genes only provide information on how to make a new
membrane from an existing one, but not on how to make a new one from
scratch. Our membranes are direct descendants of the first cell

So, I'd say that the only true replicators are complete cells which are
a combination of structural elements (cytoskeleton, centrioles,
membranes), of processors (enzymes) and which carry along the code to
make all these things, whether or not a mould is needed (like for the
membrane). By using 'replicator' for a gene, we get a confusion between
encoded genetic information and true replicators (cells).

There is another reason why I have difficulties naming genes
replicators. It is largely overlooked that only a few genes have to do
with replication. Only a few genes encode for gyrases, polymerases,
ribosomal proteins, helicase etc., i.e. for enzymes which have to do
with replication. Most enzymes have to do with cell metabolism and thus
their genes would never enable replication of the DNA-code.
Furthermore, without the information coming from these metabolic genes
there would be no replication neither, because of lack of energy and
building blocks to do so. And even in case the replicative genes could
manage on their own, there are so many of them needed that we cannot
speak of a single gene as a replicator. Genes have been arranged in
groups long time ago and they usually keep travelling along. In bacteria
this is most obvious since usually the complete genome is duplicated
exactly into the new cell, without rearrangement of the genes. They all
go together and one can speak of a genome lineage. It would be difficult
to break up these genes, since they form a community with all members
providing information for how to make a cooperative cellular system.
Even in recombinationial reproduction, all the genes are needed and only
alleles are interchangeable.

So, the idea of single genes having some meaning on their own, is rather
odd. Dawkins made the comparison with a boat of rowers working together,
but many of us deal with the idea of a gene as the replicator as if
these can easily be considered on their own.

The vehicle or interactor and the extended phenotype.
We usually speak of organisms as the phenotype of genes. I'd say that
they are extended phenotype (a term Dawkins invented to describe
specific behaviours, like those influencing the behaviour of other
species or for artefacts like bird nests). I'd say that the direct
phenotype of DNA and RNA is the protein. These are processors which
build up hugue constructions, by making cells using the encoded genetic
information (and membranes) as a blueprint. We usually confuse
multicellular organisms with phenotype, and again this leads to several
misinterpretations of what the phenotype really is when trying to apply
it in memetics.

These enzymes make also artefacts. Oxygen is an example, lignin is
another one. You could call these artefacts extended phenotype as well.
And how about the new DNA strands? Aren't they artefacts of enzymatic
activity and thus phenotype? Of course DNA-strands are a special kind of
phenotypic product, because they contain encoded information, but still
they can be considered as phenotype or artefact. (If you ask me: that is
how RNA and DNA started first: as an artefact of already complex living

The interactor: When studying cell biology, one observes that there is a
long chain of interactions going from the genome of one cell to the
genome of another cell, crossing membrane borders, and so on the next
cell, with different cells at the same time. Who is interacting: the
cells, the genes, the enzymes, the intercellular transmitter molecules?
The same is true for discussions: are they the interaction of persons,
neurons, ideas, or of a bit of all of them? Again, the interactor is
some term which comes from looking at multicellular organisms only and
making generalisations out of a special case (most living organisms are

The vehicle: we have shown above that cells are more than temporary
vehicles. While the vehicle is always presented as something transient
and temporary, it is often forgotten that physical genes are just as
Well, again this comes from considering multicellular organisms. Indeed,
they could be considered for the largest part as vehicles, but not those
cells of their germ line which will make it to the next generation (and
not all the cells taking part in budding or vegetative reproduction).

The intrinsic contradictions and uselessness of these terms should be
apparent from what I've tried to explain above.

So, many confusions exist: we speak of organisms as the phenotype, while
one should consider proteins and DNA-strands as phenotypic expressions
as well, we speak of genes as replicators while I can see only cells as
replicators and genes as something which is replicated, we speak of
phenotype and genotype while in biology they seem to be intrinsic parts
of a continuous cycle whereby it is impossible to tell one from another.
There is no way to get out of it. I suggest to use other terms, which
are more related to information theory and cybernetics and which really
say what biology - as we know it exists since at least 3 billion years -
is about.

It might be better to speak of encoded information (as in physical
genes) and of processors and/or processed information. Of course this is
not as fancy as 'replicator' which seems to be something of its own and
which has something threatening.

Mario Vaneechoutte
Laboratory Bacteriology & Virology
Blok A, De Pintelaan 185
University Hospital Ghent
Belgium 9000 Ghent
Tel: +32 9 240 36 92
Fax: +32 9 240 36 59
Editor J. Memetics:
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)