By meta-evolution, we mean the evolution of mechanisms that assist evolution. These changes can be thought of as optimizing the evolutionary algorithm. Once we adopt the view of Evolution as Algorithm we can unify understanding of a number of events in human history and ancestry that are generally studied in isolation and often not even understood as being instances of a more general pattern.

Since evolution is an optimization algorithm it makes sense to evaluate meta-evolutionary change by how it optimizes the performance of evolution according to the same criteria by which we evaluate artificial optimization algorithms: efficiency and generality. Efficiency is the speed by which the optimum is discovered. One of the hallmarks of a meta-evolutionary innovation is an increase in the speed of evolution. Generality refers to class of problems which the algorithm is applicable to. The Fitness landscape contains local peaks, so evolution faces a Global optimization problem. The power of the evolutionary algorithm is defined by its ability to find new peaks in the fitness landscape. Meta-evolutionary advances are also associated with spreading into new evolutionary niches, filling the world with life.

We identify these major meta-evolutionary events:

The origin of life itself can be seen as a meta-evolutionary event, since before then there was no evolution at all. It is likely that life did not spring forth with the slick gene / RNA / DNA / ribosome architecture that is now universal. Arriving at the point of a recognizable bacterial cell required a great deal of meta-evolutionary change, so much so that the now-lost history of origin is best understood as primarily meta-evolutionary, though there was undoubtedly also metabolic innovation associated with radiation into new niches.

Meta-evolutionary innovation is also the defining feature of the top level of classification of life: the distinction between kingdoms of eukaryote and prokaryote. Because the eukaryote nucleus can organize much more DNA, this change was crucial in enabling the evolution of more complex organisms, including all animals.

sex is the most important meta-evolutionary innovation in biological evolution. The power of sex is its ability to mix-and-match successful organisms with the reasonable expectation of creating a viable new organism. Before sex, the only mechanism of change was random mutation, which is usually harmful, and can at best only create a slightly different child organism. Sex greatly increased both the speed and power of evolution. Evolution is faster with sex because much more variation can be created without an excess of non-viable children harming reproductive success. Sex also allows the discovery of new peaks in the fitness landscape because when the children “parachute into” intermediate points in the design space they may land near a new peak.

Though bacteria have various interesting sex lives, only in eukaryotes do we see what we normally think of as sex, with 50%/50% mixing of genes, coupled to reproduction. An animal's body, or even a tree, is primarily made up of non-reproductive cells. In order to have these sorts of specialized (somatic) tissues that we and other animals rely on, it is necessary for sex to be linked to reproduction (in germline organs). Complex organisms may reproduce without having sex (parthenogenesis), but these organisms never could have evolved without sexually reproducing ancestors, and if they reproduce exclusively by parthenogenesis, then they are stuck in an evolutionary dead end.

Sex is also a clear example of the inadequacy of current evolutionary theory to explain meta-evolutionary change (see the analysis page.)

The evolution of homnids capable of cultural evolution was a major meta-evolutionary milestone, kicking off the whole process of Genetic-Cultural Coevolution. If sex creates a challenge for the selfish_gene theory, then cultural evolution is much worse. Some evolutionists offered up meme_theory, a straightforward (and inadequate) generalization of selfish gene theory to cultural evolution. Others argue that cultural change is so different that it makes no sense to call it evolution.

There was another meta-evolutionary change deep in prehistory, associated with a sudden increase the rate of cultural change, as shown by the diversification in archeological artifacts. It seems likely that there are correlated genetic changes, the product of Genetic-Cultural Coevolution.

Yet another turning point was at the onset of history, with the roughly coincident development of the state, grain agriculture and writing. This created a strong selection pressure on such civilized humans, and is likely one of the strongest drivers of recent human genetic evolution. In this process, humans have domesticated themselves, almost certainly with considerable change in human behavioral instincts. This view is in some conflict with the usual understanding of the Environment of evolutionary adaptedness in Evolutionary Psychology, which is generally assumed to correspond to tribal living more than 12,000 years ago.

Things have changed yet again in modern times (for some value of modern.) There was clearly a major innovation in both rate of change and of radiation into new niches that occurred around the time of the industrial revolution. It is mostly a matter of taste whether to identify multiple waves of change (as in The Third Wave) or a single process of exponential speeding and intensification, but there is no denying that we live in interesting times.

From the viewpoint of meta-evolution we can clarify some common misunderstandings about human evolution. There are three different mechanisms of human evolution: genetic mutation, sexual reassortment, and cultural evolution. Each builds on the forms that came before, but does not replace or stop the other forms of evolution.

Behavior drives evolution, which means that when an organism behaves in a way that changes the environment it lives in, this creates a selection pressure that drives genetic change. This is a standard idea in in evolutionary theory (associated with Ernst W. Mayr) that applies to any organism that has behavior, but is clearly relevant to humans, given the large changes we have caused in our own environment. When cultural change drives behavioral change, then culture can drive genetics. This is what Genetic-Cultural Coevolution is all about.

One thing that creates confusion is the general failure to distinguish between sexual genetic reassortment and genetic mutation, lumping both as “genetic evolution.” We believe that two serious scientific problems underlie this:

• Theoretically, it has proven difficult to show how the almost universal prevalence of sex makes any evolutionary sense (see Evolution of sexual reproduction.) Possibly because of this embarrassment, there is a tendency to minimize the importance of sexual reassortment, and to neglect its study.
• Practically, DNA technology has greatly transformed the methods of understanding evolution, allowing a much greater rigor and precision. But these methods can only examine the mutation of individual genes, and the flow of individual genes through populations. It has been possible to do interesting science without really understanding what genes do, but to understand the importance of sexual reassortment we must not only understand how individual genes work, but also how a constellation of related genes and their non-genetic regulatory sequences interact to create structure and behavior. This is a vastly more difficult problem.

Unfortunately, these difficulties combine to create an If all you have is a hammer... problem. We lack the tools to understand the importance of sexual reassortment, so the more tractable problem of genetic mutation gets most of the attention. Although there is good evidence that a few genetic mutations arose as a consequence of agriculture (such as the ability to digest milk as a adult) it is undoubtedly true that, due to its greater speed, sexual reassortment is the primary mechanism of recent genetic evolution. Yet this change is largely invisible using current tools.

Note that the general pattern is that the later forms of evolution are faster than the earlier forms, but the ultimate range of adaptation is limited by the materials bequeathed to us by the earlier forms of evolution. So cultural evolution is limited by our genetic makeup, including our gene assortments and regulatory sequences. And the power of sexual reassortment is in turn limited by the palette of genes available to work with. Rather than cultural evolution superseding sexual reassortment, or sexual reassortment superseding genetic mutation, each builds on the layer below, and often tends to increase the selective pressure, speeding up the evolution of its component parts.

If one is inclined toward teleological explanations then it is easy to see the hand of God in the theoretically inexplicable and oh-so-convenient appearance of these evolutionary mechanisms that are essential to our human state. We nave no need for that hypothesis. These mechanisms manifestly have proved themselves stable now that they are in place, so must be adaptive even in the narrow selfish-gene view. If it is difficult to develop plausible situations in which these mechanisms evolve by ordinary gradual micro-evolution, then we are forced to fall back to the position that improbable non-normative conditions led to an environment where evolution could proceed in the “correct” direction. This is not quite the same as supposing there was a Hopeful Monster that arose by mutation in a single generation, but close. Insofar as we conclude that meta-evolutionary innovations were improbable we give weight to the belief in Evolutionary contingency, but are not left entirely without a natural explanation. In particular, if we can convincingly argue that a meta-evolutionary innovation was necessary for intelligent life to arise, then we can use the Anthropic principle to turn the Argument from Design on its head. Although they have mainly been applied to cosmology, anthropic arguments provide a subtle tautological explanation of why we might find ourselves in the best of all possible universes. See also silent_witnesses and retrospective bias. Technical note: Additive Genetic Interaction.