MEMEPOOL VERSUS GENEPOOL

In analogy with genes, memes are presented as floating in a memepool (Dawkins, 1989 page 192-194). However, it is often unclear what a memepool is exactly, if it has boundaries, and in what respect memes disseminating through a memepool are and are not analogous to genes disseminating through a genepool. To achieve insight into what a memepool is I shall firstly investigate what a genepool is, and then elaborate on a possible analogy. While doing so I will also elaborate on some analogies and disanalogies.

Genepool versus population, or replicator pool versus unit-pool

First of all, in my view the genepool [4] is a statistical description in terms of relative gene-frequencies, and must be distinguished from a sexually reproducing population of organisms. Instead it is a mental construct to describe genetic changes in a population. It is a bookkeeping device in which all the genes from all the organisms considered to be in the population are accounted for. It is the boundary over which no genetic communication takes place of the actual population that forms the shores of the pool. If genes can be said to travel through anything, it is through the population, and not through the genepool. If the relative frequency from a specific gene increases, this gene can be said to spread through the population. Notice that 'spreading' means a change in relative numbers of organisms, often accounted over more generations.

In terms of general replicator concepts, a replicator pool is a content description for a unit (population of organisms in the case of genes as replicators) in which communication or transmission by replication takes place. The word replicator pool (genepool) denotes to the replicator contents of this unit (the genes in the population). From now on, I will refer to this unit as unit-pool (population). The boundary of the unit-pool determines the 'boundary' of the replicator pool (what replicators are counted in the replicator pool), this can be described as a border which replicators do not cross. The consequence is that particular replicator lineages run only within a unit-pool.

In a population of sexual reproducing organisms, replicators are part of lineages that meet each other in vehicles (organisms) in which they are replicated at the same time, by the same replicator mechanism(s). The borders of the population are determined by the fact that replicators, which might exist within other populations are not replicated by the same replication mechanisms. Stated in terms of replication mechanisms, we can say that every replicator mechanism, and the entity it is in (the organism or vehicle), that replicates a replicator of a specific lineage, belongs to the same unit-pool (population) as every other mechanism that replicates a replicator of the same lineage. Within a population different replicator lineages can be distinguished because of splitting processes like meiosis (recombination). In successive generations, lineages that were parted before can meet again in a vehicle in which they are replicated together again. For instance in sexual reproduction, a replicator like the y-chromosome can be distinguished that is replicated as a whole to the progeny, and thus is a lineage (unbiased by recombination mechanisms). If the y-chromosome is in the same organism as an x-chromosome, it won't be with it or its descendants again in the next generation (if there are only two sex-chromosomes in one organism, and every parent donates one). In the next generation however, the y-chromosome can end up in the same organism again with one of the descendants of the x-chromosome.

In biology, it makes little difference to say at any point in time that the genepool consists of all genes, or to think about it as a group of physical bodies (the organisms) genes travel through because genes and organisms go together. Genes mostly 'travel through' the unit-pool along with the organisms they build, via sperms or eggs Dawkins (1989, page 192). In this way 'memes travelling through a memepool' becomes analogical: Dawkins sees memes disseminating through (a part of) society. His example of schools featuring the 'craze' shows this. In such examples a meme spreads through units that make up a society (schools, or better the individuals in such schools). For example the 'craze' can first appear at one school, then at another one, when individuals from different schools communicate it, and so on.

This dissemination of memes through society looks superficially analogous to genes spreading through a genepool. However, as stated above, a replicator pool is no physical entity, the unit-pool is. Dawkins (1989) doesn't make a difference between the unit-pool, and the mental construct to describe this pool (the replicator pool). He describes the genepool as something physical through which genes travel, for instance in the sentence 'Just as genes propagate themselves in the genepool by leaping from body to body via sperms or eggs, so memes propagate themselves in the memepool by leaping from brain to brain via a process... (p 192), or 'those genes that behave in such a way as to increase their numbers in future gene pool' (p196), and 'the old gene-selected evolution, by making brains, provided the soup in which the first memes arose' (p194).
Assuming that Dawkins knows the difference between a unit-pool and a replicator pool, his choice of language is unfortunate, because it gives food to the thought that a genepool is something physical through which things can travel.

Virus versus organism-genes and memes

Even if we grant that Dawkins' way of using the concept genepool is just a figure of speech, there are still problems with his view. Genes in the genepool are replicators that are involved in building the vehicles (organisms) they reside in. Memes do not, they only reside in vehicles that are built by genes (humans) as well as in other vehicles (like books, etc.). As shown above, in a unit-pool, the vehicles built by the genes are the physical entities with which genes 'travel'. So genes do not travel from organism to organism, but they are recombined from previous generations and build new organisms that make up the next generation.

Above I have carefully avoided the cases where genes do jump from organism to organism to avoid a too complicated picture. But such genes do exist, in viruses, plasmids, and other examples. In fact, if we look at the definition I have given of the boundaries of a unit-pool, viruses belong to that pool, since their replicators (constituting distinguishable lineages) meet replicators that belong to the hosts' lineages in the same cells where they are replicated by the same mechanism (however, not always simultaneously).

In the analogy Dawkins sketches for memes he explicitly uses the examples of vertical spreading genes through the population, in the form of viruses (versus the vertical travelling that characterizes genes that are involved in building the organisms). Boulding ( 1978 ) and Boyd and Richerson (1985) call this 'jumping' from entity to entity in social evolution multi-parental transmission, while Campbell (1965) refers to this as cross-lineage borrowing in social evolution. With this virus-meme analogy Dawkins also introduces the concept of a parasitical relation into the gene-meme analogy. In endnote two I shall deal further with this concept, and the way Dawkins uses it [II] .

The virus analogy is probably better than the analogy with genes that build and travel with organisms through their whole life cycle. Like viruses, memes do not build organisms, or are alive in the sense that they have metabolisms, but they can evolve by natural (or intentional) selection (Hull, 1988a ). This analogy of viruses disseminating in a hostpopulation and memes disseminating through society is appropriate, but the use of the word genepool for the unit-pool of the hosts is unfortunate, as also the claim that memes must be parasitical towards their hosts. In the case of the virus, according to the definition above, it can be said that they indeed belong to the host genepool, but this definition is not a common one. Probably most biologists will not view viruses spreading through a host population as genes that spread through a genepool. The mixing of two metaphors, virus/host and genes/genepool, will only be confusing, also for biologists. Memes spreading in a unit-pool are thus analogous to viruses in A) that they do not 'build' the organisms they 'travel through' and B) their horizontal spread. They are analogous to organism-genes in that they can be beneficial for the 'host' (see endnote 2 ).

Memes differ from both virus-genes and organism-genes because A) they are replicated in brains, and not by DNA replication mechanisms, and B) they do not have to be instructional toward (a part of) the behaviour of the host. Memes are analogous to both virus-genes and organism-genes in that they are replicators that can evolve.

A definition for the memepool and its unit-pool

After the analysis of some analogies and differences between memes, genes and viruses we can return to the definition of a replicator pool, and unit-pool in memetics. According to the replicator pool characterization previously, the memepool is a content description (in terms of memes) of a unit-pool which boundaries are determined by communication or replication mechanisms in it that replicate replicators from identical lineages. A memetic unit-pool thus exists of replicator mechanisms (brains, or minds), that are connected by replicator-lineages. According to this definition, two minds that share a meme-lineage belong to the same unit-pool.

Boundaries between memetic unit-pools occur: where humans do not communicate, where memes are transmitted, but not replicated, because they are not understood, or ignored for other reasons. Such almost physical boundaries occur between people that speak different languages, or that use different paradigms in a scientific community.
However, if we take the criterium that all humans (being the replication mechanisms) that have ones shared, or share only one lineage must be counted as vehicles in the same unit-pool, virtually all humans belong to one big unit-pool in contemporary society. This is the case of course because of the horizontal multiple parent spreading, which characterizes a great part of memetic dissemination, as well as because there is an enormous amount of memetic information in every human.
For example, if a theory is used in a scientific discipline the replication sphere for this lineage would be the minds (or parts of the brain that replicate) of scientists involved. These scientists are thus part of the unit-pool for this theory lineage.
If we imagine that the scientists work on environmental problems, like acid rain, they will probably replicate theories that belong to the same lineage, as government workers that are involved in policy planning to solve that problem. According to the characterization of a unit-pool, both scientists and government workers should thus be included in the unit-pool. Both share identical meme lineages.
According to the definition of the unit-pool different lineages that are replicated simultaneously belong to the same replicator unit, and pool. This means that all other lineages that are replicated by the scientists must also be counted as a part of the memepool (If we relax the criterium that they must be replicated at the same time). But if the scientists also share some meme lineages with philosophers, regarding methodology for instance, and with many other professional groups, not to mention with people they deal with in the personal sphere, the unit-pool will in reality virtually be unlimited, since the philosophers, etc. will in their turn share identical meme lineages with other people again, etc. While genetic unit-pools are highly constrained because interspecies breeding is impossible, and populations of the same species can be divided geographically, memetic evolution might not have such boundaries.

The memetic unit-pool as a useful concept

A less strict definition of unit-pool in memetics may be practical with regard to the criterium that replication mechanisms belong to the same unit-pool if they only share one identical meme lineage. Something similar occurs in biology where populations are distinguished that only occasionally exchange genes. In the same way, some memes might be transmitted from one unit-pool to the other as long as this is the exception rather than the rule.

Further relaxation of the definition can be made with regard to different (classes of) memes. Two groups of people might communicate frequently about the weather for instance, but never replicate each others religiousness memes. In fact, further relaxation results in the appearance of all kinds of interesting questions:
In groups of people different memes present can be unequally dispersed; the unit-pool for one meme (meme A) can overlap with the unit-pool for another (meme B). A newly introduced meme (C) could be accepted in unit-pool A, but not in unit-pool B. Meme C can for example be in conflict with meme B, but not with meme A. It is possible that meme B and C describe rival explanations, or maybe meme A must be mastered before meme C can be taken up, etc.

In this approach we can describe situations where already 'installed' memes select new memes, or create a 'niche' for new memes, yielding memetic dispersal structures in unit-pools. Dennett ( 1991 , chapter seven) for instance, has developed a view where already installed memes are necessary for the uptake of other memes. Notice that such logic can apply to memes as ideas, but also to memes as norms, or instructions how to handle particular situations. The research on dispersal structures, as well as on memetic structures within minds (what ways of thinking must be mastered to understand differential mathematics, or evolutionary theorizing for instance) is a potentially rich source for empirical research that I would call memetic.

Above I have proposed definitions for memepool, and unit-pool. They are necessary concepts for a systematic comparison of the meme-gene analogy. I believe the characterization of the unit-pool, and its content of memes is analogically a good one. To address the phenomena of memetics, such as why memes spread, we need to go further. I have mentioned the selective environment as a key-ingredient for the explanation of genespread in a population, and also that memes can select memes. Below I will elaborate on what memetic selection can denote to.

next part

Back to hypertext essay on memes

Back to Memes Back to Evolution Back to Hans-Cees