We have defined evolution as a 'pattern of change'. This allows us to imagine different patterns. We view all these patterns as members of a large evolutionary family, which is studied by the discipline of 'Big Evolution'.

The range of patterns of evolution which are studied by Big Evolution includes for example a mirror image, where objects form combinations (e.g. the 'combinatorial evolution' of Brian Arthur). And while Darwinian evolution has problems dealing with transitions in the type of objects between generations, Big Evolution includes patterns of change that incorporate these 'major transitions'. The pattern of evolution can furthermore accomodate to nested forms of evolution, to the inclusion of life histories, to the evolution of abstract concepts, etc.

For references and detailed information about this topic:

A rethink of evolution, and the construction of a family of ‘patterns of evolution’ inside and outside biology Gerard A.J.M. Jagers op Akkerhuis, Hendrik Pieter Spijkerboer, Hans-Peter Koelewijn Academia.edu 2014


Several generalisations of the concept of evolution have been suggested, including "generalised Darwinism" (Hodgson & Knudsen 2006), "universal Darwinism" (Dawkins 1983) and the description of evolution as a substrate neutral algorithm (Dennett 1995). And in biology a more general version of the Modern Synthesis is in the making which is called the Extended Synthesis.

The approach we follow in this website shows analogies with aspects of the latter approaches. We focus, however, on the possibility of defining the concept of evolution by means of a 'pattern'. Why do we add yet another definition of evolution? To answer this question, the following paragraphs explain that 'evolution' can be described by a least complicated 'pattern of evolution'. This basic pattern can be adjusted to handle more complicated situations. We have named the scientific discipline that studies the entire 'family' of all conceivable patterns of evolution "Big Evolution".

The following text offers examples of variations on the least complicated 'pattern of evolution'.

'Confluent' and 'diffluent' patterns of evolution

confluent pattern of evolution.jpg
Figure 1: A confluent pattern of evolution. The fact that only objects A1 and A2 act as the basis for B, represents the pattern of selection.
The least complicated graph for evolution that we discussed in a preceding chapter focused on classical Darwinian evolution. A single object gives rise to several second generation objects, while these second generation objects comply with the pattern of selection. The relationship between a single first generation object, and several second generation objects, is a 'one-to-many relationship' that is described here as a ‘diffluent' pattern.

The diffluent pattern has a mirror image: the 'confluent' pattern. This mirror image has been named ‘combinatorial’ evolution by Brian Arthur (2009). In biology, a confluent pattern implies that two organisms combine and form a single unicellular or multicellular organism. It must be noted that in a confluent pattern, the pattern of selection is located before the formation of the next generation object.

In biology the confluent pattern will be produced by organisms. To know exactly when a combination of cells results in a single, new organism, we use the definition of an organism that has been proposed in Jagers op Akkerhuis (2010) and Jagers op Akkerhuis (2012) as follows: “any operator of a type that is equal or higher than that of a cellular operator”. Based on this definition, the formation of a zygote is viewed as a biological example of a confluent pattern of evolution.

Transitions within operator type (WOT) and between operator type (BOT)

It is not easy to describe a difference between first and second generation object(s). In many cases, however, this problem can be solved by by using the logic of the operator hierarchy. Based on the operators and their types, one can identify processes that don't and processes that do lead to a an operator of a new type.

Firstly, the operator type can stay the same between generations. For example a bacterium that splits produces two new cells, each being a bacterium. When the operator type does not change, we speak of a ‘within operator type transition’ (or ‘WOT’ transition). Following a within operator type transitions, the operators can still be of a specific subtypes. For example when different atoms are formed, these can be of the subtype of Helium or Argon, or of the subtype 'Lanthanides'. And the operator type of the 'multicellular organism with neural network' includes for example the subtypes of the Homo sapiens and Canis lupus. In general, there is not a single best way of assigning operators of a type to subtypes. Various (partially overlapping) kinds of subtypes can be distinguished, such as ‘noble gases’ and ‘lanthanides’, or ‘parasites’ and ‘mammalia’.

Secondly, the operator type can change between generations. If a process leads to an operator of a new type it is called a ‘between operator type transition’ (or ‘BOT’ transitions). The explanatory/causal mechanisms for the emergence of BOT transitions have been discussed e.g. by Maynard Smith and Szathmáry (1995), Blackstone (1995), Michod (1999) and Godfrey-Smith (2009). The mechanism we focus on is the emergence of closure as a structural/topological criterion.

A pattern of evolution for life histories

Life history evolution.jpg
Figure 2: Two examples of how the minimal graph for evolution can be extended to analyse the evolution of life histories. Part A: a pattern of evolution for life histories in which the life history with a multicellular extension is the most ‘fit’ option. Part B: a pattern of evolution for a neural network as part of a multicellular life history. Black double arrows: within operator type transition as part of a life history. White double arrows: between operator type transition as part of a life history. E=egg cell. Z=zygote. MC=multicellular organism. NN=neural network organism.
In Figure 3A we add a multicellular stage to a life history as an event that fits to the pattern of evolution. The viewpoint of a life history implies that the genes of the original cell code for three things: 1. the joining of the daughter cells, 2. the formation of plasma strands between these cells (plasma strands define the operator type of the ‘multicellular organism’) and 3. the morphological development of the multicellular unit that is produced in this way.

In Figure 3B we show a life cycle that reaches the neural network stage. The evolution of a neural network in a multicellular organism implies that generation after generation, multicellular organisms develop longer and longer extensions of cells because the resulting improvement of the transport of stimuli through their bodies causes selective advantage (e.g. Wiljes et al. 2010). Over many generations, cells with extensions have gradually become 'neurons'. This process can be viewed as a series of many small within operator type transitions (WOT transitions) that finally add a between operator type transition (BOT transition) to the life history. Figure 3B focuses on the outcome of this evolutionary trajectory, which is depicted as the addition of a discrete new life history stage.

Nested patterns of evolution

An example of an endosymbiont operator is a protozoa with its endosymbionts (such as mitochondria). In this situation it is possible that patterns of evolution exist at three different levels of organisation. At the highest level we have the protozoa which may divide and produce next generation protozoa. At one lower level, one can observe inside the protozoa several generations of the mitochondria. The next generations of mitochondria may comply with a pattern of evolution. Finally, if parts of the genome of the protozoa and/or the endosymbionts are replicated by their cell, during the life of the cell, this aspect will contribute to the production of many new DNA parts, which can comply with a pattern of evolution. From the viewpoint of the protozoa, the evolution of the endosymbionts and their distribution over the next generation protozoa cells can be viewed as a cause of variation of the protozoa over generations. The pattern of evolution can now be recognised to occur in a nested way: the protozoa i the highest level (and for this reason the focal level). Inside we find two additional levels: of the mitochondria, and of replicating genes in mitochondria.

The use of abstractions in a pattern of evolution

In principle, the nodes in any graph of a pattern of evolution can be physical objects (or 'tokens') as well as abstract kinds. If one uses 'kinds' one has to be very careful that, without exception, all objects comply with the kind. This demand can be met easily if one discusses atoms or molecules. If one uses the concept of the population, however, it is very difficult to actually prove that all members, without exception, represent the 'kind' of the population.

A combined pattern of evolution that matches the operator hierarchy
Operator Evolution.jpg
Figure 3: The summing up of many individual patterns of confluent and diffluent evolution (at the level of types) to a single pattern for Operator Evolution. A: Part of the pattern of Operator Evolution. B: Underlying confluent and diffluent physical processes. White dashed double arrows indicate confluent processes that lead to operators of a new operator type (confluent BOT transitions). Black dashed double arrows indicate variation based on condensation reactions (confluent WOT transitions). Thin dashed grey arrows indicate diffluence through multiplication (diffluent WOT transitions). Black dashed arrows indicate the translation of the outcome of physical processes to a type-based approach.

One can combine the above patterns for the creation of a pattern of evolution that describes the entire operator hierarchy. Each individual pattern of evolution in Figure 3 complies with the demands of the least complex graph for evolution.

Confluent patterns of evolution are used to describe how within operator type transitions lead to different kids of atoms and molecules (left side of part B in Figure 3).

Diffluent patterns are used to describe the formation of various kinds of single celled (bacterial) operators and of (single celled) endosymbiont operators (Figure 3 only shows part of the entire operator hierarchy).

Confluent and diffluent patterns of evolution are combined in the sequence of type-based patterns of evolution which is typical for the operator hierarchy. If one uses this logic, this allows one to underpin the statement that the operator hierarchy can be seen as an evolutionary pattern.

The science of Big Evolution

Above we have discussed a range of patterns of evolution. the selected patterns will most likely represent only a part of all possible patterns of evolution. All these patterns of evolution can be seen as one large 'family'. The study of all patterns of evolution in the context of their membership of the family of all possible patterns of evolution is what we define here as the science of "Big Evolution".