Self-Directed Evolution

Self-Directed Evolution

Hi, I'm Tim Tyler - and today I will be discussing the idea of Self-Directed Evolution - a phrase which refers to the concept of biologial evolution being directed by intelligent agents.

When considering the ways in which evolution may be directed it is useful to consider a simple symbolic diagram of the evolutionary process:

Here, the processes responsible for generating variations in inherited genetic information are represented on the left, and the processes responsible for deciding which variants persist are represented on the right.

When intelligence affects the processes responsible for generating variations, that is a type of directed mutation which we will call an intelligent mutation.

When intelligence affects the processes responsible for deciding which variants persist, we will call that "intelligent selection".

Intelligent mutations are a phenomenon that has mostly been associated with cultural evolution so far - though there is a case for it being implicated in DNA evolution via the Baldwin effect.

Intelligent selection has been a significant factor in evolution so far:

It is how sexual selection in higher animals works. Female choice typically involves females getting males to display for them - and then using their intelligences to select between them.

Natural selection can also involve selection by intelligent agents on other members of the same species. The most obvious example is probably male combat. Males sometimes kill each other - and often have to decide when to run and when to fight.

Lastly, there is the Baldwin effect. In this, behaviour patterns repeatedly acquired by organisms in their own lifetimes can alter the way in which selection acts on them. An example would be birds that learn to open milk-bottle tops. This behaviour is learned - and can be transmitted culturally - from bird to bird. Once the behaviour exists, birds are then subject to selection, favouring lactose tolerance - and other traits related to bottle-opening and milk-drinking. Here is a case of a psychological disposition going on to result in subsequent germ-line ramifications.

However, even if decisions about who dies and who reproduces are often being made by intelligent agents, the overall behaviour of the resulting system may not necessarily appear to be very smart. Similarly, a crowd may be composed of many intelligent agents - but its overall behaviour may not necessarily seem to be very smart either. So, the question arises of whether it is possible for the whole evolutionary process to act more like a single intelligent agent.

We know that organisms could be made effectively immortal - and also that self-improving systems could be made which progressively modify their own structure in order to grow and develop.

In the future, there is the possibility that a new mode of evolution will emerge - in which a single, large, intelligent agent is constructed which is capable of controlling its own evolution.

Such an organism would have no competitors or peers. It would not senesce or die - rather it would continually reinvent itself.

In this kind of scenario, the organism would not necessarily pursue the same goals that traditional biological systems do. Natural selection tends to produce systems which degrade the available sources of order as rapidly as possible - by using the available resources to make copies of their own genomes. However, if a single large intelligent agent were in charge of the direction of evolution, it could conceivably select other goals to pursue.


So, we can imagine an evolving system which is as completely in charge of its own evolution as it is possible to be. The next question is whether such systems will ever arise. I fully expect that they will. Over time, nature has built progressively larger cooperative systems. Cooperative systems do very well competitively: if you look at multicellularity and the social insects, these developments have been wildly successful. Also, now, genetic engineering has destroyed the barriers to gene flow between different species - so now all living things can more easily be united - and can collectively utilise the discoveries of all the organisms on the planet.

We are currently witnessing the emergence of planetary-scale cooperation via the internet. It seems to me that the process of globalisation is likely to lead to a world government, a world economy, a world currency, a world language, a world immune system, and a world mind, which will then effectively take control of much of the evolution of the biosphere.

Initially, there may still be some quarrels within the system - due to internal genetic conflicts of interests. However, it seems likely that such conflicts will be quelled - and eventually the creature's insides will come to operate harmoniously.

Another question is whether such a cooperative entity could survive as a unified organism - even if it spreads out across multiple star systems, or across multiple galaxies. It seems to me that fragmentation could - and probably would - be avoided.

If such an entity dissolved itself, then another would probably form. Eventually, a stable one would arise, and then that is what would be observed from then on.

Another question involves what aims such a system might have. This is a non-trivial question. My guess would be that it would probably have much the same aims as the evolved system that preceded it - so it too would be inclined to use the available resources to rapidly expand its domain and thus ensure its own survival. However, it seems possible to imagine systems being built with other drives and goals.

Then there is the question of whether such a system is desirable, or avoidable. Already we can see an anti-globalisation movement - so it seems obvious that not everyone will like the idea. Is there anything that can be done to prevent it? Realistically, I doubt it. I see the situation as being essentially another Tech vs Luddite battle - and the Luddite factions rarely seem to get enough power to actually offer much resistance.

What about capitalism? Today we have capitalist political systems that attempt to preserve competition between companies using a monopolies and mergers commission. However, the system is clearly a dreadful one - and the government only uses it because it hasn't yet come up with a superior alternative. Also, note that - as in the proverb - each government has only one monopolies and mergers commission.

What about democracy? - yes, there are mock battles between factions for control of government, but - as in the Yes Minister TV show, much of the government infrastructure is unaffected by such reshuffles. Democracy seems unlikely to prevent a single large governing body forming - but might influence how it is run.


The dynamics of such systems can be seen through the lens of today's politics - as being associated with the battle between capitalism and socialism.

My expectation is that capitalism will be discarded as a stupid, inefficient and primitive system - and future metabolic activities, will be managed and run directly by the government.

Another way of looking at the topic is in the context of the battle between individual liberty and the state.

My expectation is that the state will obliterate individual liberty - and that any autonomous individuals will become slaves of the state - much like worker ants are slaves to the hive mind.

My expectation is that the march towards globalisation and cooperation will continue until all traces of conflict between sentient organisms have been eradicated. There may still be viruses - and other conflict on a small scale - but it seems likely that even that kind of activity will be heavily subdued.

Of course, there will still be trial-and-error exploration. However, the trials and errors will probably not involve death or damage to sentient agents - or a battle for reproductive success between them. Indeed, much trial and error will probably be done under simulation.


In conclusion, Self-Directed Evolution looks likely to represent the future of all living systems on the planet. However, so far there has been relatively little exploration of the dynamics of such systems. Since this seems likely to be the shape of biological evolution in the future, it seems to me that some study is warranted.


  1. One Big Organism - Will all life unite?;
  2. Evolution Sees! - Evolution is not "blind";

Tim Tyler | Contact |