I suggest to approach the problem of increasing complexity very differently.
Many of the arguments above with regard to fitness (absolute or relative) seem not able to
explain the increase of complexity. One can agree that a system which can exist in
environments A, B, C, D and E is probably more complex than systems which can survive
only A, B, C or only D, E, F. But is it more fit than a system which e.g. can only survive in
environment F? In case this environment F is the most likely one to be found on Earth, it
certainly is not. Take a nuclear war which can be survived by some bacteria only. This is only
an extreme example to make the case clear, but it exemplifies a generally applicable remark.
Complexity is no guarantee for fitness, which - I think - is always related to the environment.
Furthermore, the 'Red Queen Principle' (or the arms race principle) shows how organisms
become more complex while remaining equally fit.
Here is how I propose to look at it:
1. One could state that there is no direction of evolution with regard to the increase of
complexity of individual living organisms. Evolution leads mostly to equally complex,
sometimes to less complex, and rarely to more complex living organisms.
2. Equal complexity is the rule: Evolution of bacteria leads to other bacterial species or other
bacterial indivduals. Evolution of mammals usually leads to other mammals.
3. Evolution towards less complexity is often observed: Mitochondria can be considered as
symbiontic bacteria which do not need to be as complex as the free living, more original
bacteria, and indeed mitochondria are less complex. Similarly, many parasites are far less
complex than their free living ancestors.
4. Evolution towards more complexity actually is statistically very unlikely. The first
unicellular organisms, the origin of photosynthesis, the origin of the eukaryotic cell are only
possible through long chains of unlikely events. However, since statistics is about large
numbers and since evolution deals with large numbers of reproducing entities, very unlikely
events may become very unlikely to unevitable.
There is still another aspect about more complex organisms: they may conquer new niches,
not occupied by any other organism. This is the empty niche which is always open to more
complex organisms (multicellulars can feed on unicellulars, which could not feed on
multicellulars. Only later, unicellulars could parasitize on multicellulars). Therefore, more
complex forms should not be explained solely by the fact that they are more fit than their
competitors, but especially because they always find empty niches with lots of possibilities
before them. This is confirmed by the fact that we see such rapid radiations after new
inventions: there is so much space, food, energy, ... available and no competition for it, that
almost every concept at the new complexity level will work. Take pentaradial organisms: they
represent a quite strange concept of bodily organisation, yet they exist and still are very
successful (Echinodermates, like seastars etc.).
Another possible confirmation of the importance of the empty niche in explaining how
evolution towards more complex organisation is possible, comes from the observation that -
soon after they came into existence - multicellular animals rapidly radiated into all presently
known (and many already extincted) animal phyla. No new phyla arised thereafter, because
very soon there was too much competition among the multicellulars. Once the niches were
occupied, no new forms could arise.
A similar example is maybe seen with new ways of thinking which initially lead to a rapid
radiation of all possible kinds of new related ideas. Take as an example memetics in its
current status of protoscience.
5. In summary, the first nucleotide encoded, cellular living organisms are still here and still
crucial to any form of current life. From this complexity level onwards there is always
evolution possible towards less or more complex forms of organisation. However, while less
complex organisation is usually a special adaptation to some very specific niche (e.g.
parasitism), the more complex forms of organisation will usually give raise to a multitude of
other forms of life at this new level of complexity because of the high probability of
encountering empty niches. One could say that evolution seems to be open ended towards
more complexity. Once this new level of complexity is well established, high numbers of
organisms exist, many new interactions have occurred (also with less complex organisms),
all niches are occupied, and it becomes possible that the event of real innovation repeats
itself, since at this moment highly unlikely events become almost unevitable (because of the
law of numbers), giving raise to a next level of higher complexity.
6. Thus, although complexity of part of the individual systems has increased during
evolution, this cannot be explained by the principle of fitness, but can be regarded at as the
consequence of a 'higher order' principle.
We' better consider the process of life as a whole, instead of looking at individual
subprocesses (organisms). Since evolution leads to more diversity of organisms - through
multiplication of organisms in combination with mutation and recombination of the genetic
code, the interactions between always more of always more diverse and possibly more
complex organisms - which can be considered as the temporary subprocesses of the process
of life - can possibly become more complex. The process of life becomes more complex and
only now and then this leads to more complex subsystems, adding substantially to the
complexity of life.
Also remark that the methods of information exchange between the different organisms
obviously become more diverse. From chemical interaction only at the cellular level, to
additional interaction by sound and visual symbols at the animal level, to additional
interaction by representative sounds (spoken words) by humans, to additional interaction by
means of visual representative symbols (written words) by recent humans and science.
Interaction by means of digitalized, electronically encoded information exchange between
humans and computers is of course the most recent new development