Informational Genetics

Informational Genetics

Informational genetics refers to the unification of Shannon's information theory and genetics.


The unification of information theory and genetics was pioneered by G. C. Williams. Some quotes ilustrate his perspective:

"In this book I use the term gene to mean 'that which segregates and recombines with appreciable frequency'" - Williams, 1966, page 241.

"In evolutionary theory, a gene could be defined as any hereditary information for which there is a favorable or unfavorable selection bias equal to several or many times the rate of endogenous change" - Williams 1966, page 25.

"A gene is not a DNA molecule; it is the transcribable information coded by the molecule" - Williams 1992, page 11.

This view of genes as informational entities was subsequently embraced by numerous theoretical biologists - and popularised by writers such as Dawkins and Hull.


The specific definitions which Williams gives have some problems.

Defining a gene (or, strictly speaking, an allele) as something which has an "appreciable frequency" results in novel mutations which occur only once in the population being denied the status of genes.

Defining genes (or, strictly speaking, alleles) in terms of the selection pressure on them doesn't seem right either. If an allele happens to be neutral, surely it doesn't suddenly stop being an allele.

Worse, these definitions conflict with each other. One mentions selection, and the other mentions frequency. An allele can rise in frequency for other reasons besides its selective benefit - e.g. because of linkage to beneficial alleles present at other loci.

The reason Williams adopted these definitions appears to be because he wanted his definition to address the issue of how big genes are.

Molecular geneticists sometimes define evolution in terms of changes in allele frequencies. With such a definition, the issue of how big alleles are can seem to be important.

However, these definitions introduce serious problems - so serious that we must reject them as being terminally flawed.

Alleles should broadly map onto the idea of Mendelian factors. Neither the strength of the selection pressure on them nor their frequencies are relevant to this.


The key insight of Williams - that genes are informational - is very important.

At first glance some may regard this idea as trivially wrong. Organisms inherit more than just information from their ancestors. They inherit geographical location, their ecosystem, pathogens, grandfather clocks, traditions - and all manner of other things.

However, if you look across multiple generations and ask what actually persists over extended periods of time, the answer is almost always some kind of digital information.

Grandfather clocks disintegrate, but digital patterns can live forever. Digital information can be backed up and copied. It is therefore potentially immortal - it can evade entropy and death. No physical object can do the same. What evolves and changes over time is, at its base, informational in nature.

Phenotypes may appear to persist overtime - however, they do so as a consequence of the persistence of genetic information.


Dawkins originally used the term "gene" to refer to these persistent informational entities. However, after a while, he came to regard the term "gene" as too heavily overloaded - and so adopted the term "replicator" instead.

David Hull also began to use the term "replicator" in this context.

Hull defined a replicator as: "an entity that passes on its structure largely intact in successive generations" - Hull, 1988.

Unfortunately, the actual condition for an information carrying substrate to support an evolutionary process is that it is capable of sustaining information transfer across multiple generations. How much of its "structure" remains "largely intact" in the process is not critical. Self-encrypting computer viruses demonstrate one of the problems associated with Hull's definition.

The "replicator" terminology has several problems:

The term replicate has the implication of high-fidelity copying. However, we know from information theory that information can survive in relatively poor-fidelity copying systems - provided they employ redundancy and error correction. This insight dates back to J. von Neumann's paper: "Probabilistic logics and synthesis of reliable organisms from unreliable components."

Also, the study of genes is genetics. And a collection of genes is a genome. What is the study of replicators? And what do we call a collection of replicators?

Adopting the replicator terminology requires us to reinvent parallel copies of existing terminology. We see this has already happened in terminology associated with cultural evolution. Memetics is the study of memes. A memeplex is a collection of memes. Sue Blackmore has proposed that we need additional concepts, temes and presumably temetics - to deal with inheritance mediated via machines. Should we also have renetics, the study of replicators - and a rene - an individual replicating entity?

In my view, this is all poor terminology. We have only one fundamental entity here, and do not need multiple names for it. The basic concept is that of inheritance.

The best way forwards is to abandon the replicator terminology - and go back to the original vision of Williams and Dawkins - by using the terms "gene", "genetics" and "genome".

Yes, these terms are overloaded, and yes, that is a potential cause of confusion - but using them really is the best way forwards for evolutionary biology.

Definition of Gene

Since Williams did not provide a viable definition of "gene" I propose that a gene be defined as being "a small section of heritable information".

This avoids the problems the definitons of Williams - while dodging the question of how big a gene is - an issue which I do not see any way to address concisely enough to go into a short and snappy definition.


The informational definiton of a gene faces another issue: it makes it very clear that it is impossible to evade the conclusion that human culture constitutes heritable information.

If you define genes as small bits of heritable information, then ideas - and other cultural information - inevitably fits the description.

Culture becomes a type of genetic information which simply happens to be inherited in media other than DNA - and is passed down the generation via pathways which are largely independent of DNA-based inheritance.

Conventional discussions of whether a trait is genetic or cultural turn into incomprehensible nonsense: culture is genetic - by definition.

Some people's reaction to this is to reject this redefinition of the term "gene" as flying in the face of common usage - and common sense.

However, in my view, the proper reaction is different. The idea that culture is genetic information which happens not to be represented as DNA represents a shift in perspective which contains deep insights.

Cultural evolution is not merely a process analogous to DNA evolution. It is part of evolution. The new replicators are not merely analogous to genes. They are genes. Companies, religions, books, movies, etc do not just evolve and change like living entities, they are living entities. Culture becomes part of biology. Memes are simply a type of gene that is not made of DNA.

Artificial Life enthusiasts have long derided definitions of life, genes and biology that reference the details of biochemistry - as carbon chauvanism. The belief that genes are necessarily made of DNA is castigated as being nucleic-acid centrism. Such ideas represent out-dated thinking, which biologists have got to get over.

Cultural evolution

This absorption of cultural evolution into the mainstream of evolutionary theory is long overdue.

As Dennett puts it:

Many Darwinians are anxious, a little uneasy, would like to see some limits on just how far the Darwinism goes.

It's all right, you know. Spider webs? Sure, they are products of evolution. The World Wide Web? Not so sure. Beaver dams, yes; Hoover Dam, no.

What do they think it is that prevents the products of human ingenuity from being themselves fruits of the tree of life - and hence in some sense obeying evolutionary rules?


Cultural evolution is evolution - according to modern definitions of evolution:

Biological Evolution entails inherited changes in populations of organisms, over a period of time, that lead to differences among them.

- Monroe Strickberger, Evolution, 1996.

Nowhere does it say that inheritance must take place via DNA.

Strickberger goes on to put the case for cultural inheritance concisely:

In short, humans have two unique hereditary systems. One is the genetic system that transfers biological information from biological parent to offspring in the form of genes and chromosomes. The other is the extragenetic system that transfers cultural information from speaker to listener, from writer to reader, from performer to spectator, and forms our cultural heritage.

- Monroe Strickberger, Evolution, 1996.

If it fails to embrace cultural evolution, evolutionary theory fails to account for the origin and development of mankind - and also, of course, the development of modern science and technology. That would be a miserable failure - and a totally unnecessary one.


  1. G. C. Williams - Adaptation and Natural Selection - 1966

  2. G. C. Williams - Natural Selection: Domains, Levels, and Challenges - 1992

  3. Richard Dawkins - The Selfish Gene - 1976

  4. David Hull - Science as Process: An Evolutionary Account of the Social and Conceptual Development of Science - 1988

  5. Monroe Strickberger - Evolution - 1996

Tim Tyler | Contact |