Irreducibly Complex systems are those systems (man-made or otherwise), where removal of critical core parts results in malfunction.
By way of contrast, fault tolerant systems allow removal of parts or entire sub-systems, yet intended function is still retained. Removable parts or subsystems in fault tolerant architectures are also contrasted with useless parts which serve no purpose. Like spare tires, removable parts in a fault tolerant systems can still serve a purpose even if never used.
From Wiki on Fault Tolerant Intelligent Design:
In engineering, fault-tolerant design is a design that enables a system to continue its intended operation, possibly at a reduced level, rather than failing completely, when some part of the system fails.
A fault tolerant system can be composed of several irreducibly complex systems. For example, the space shuttle has 5 navigation systems each capable of serving as a sufficient navigation system in case of damage or failure of the other 4. Many man-made mission critical systems have fault tolerance, redundancy and margins of safety built in.
Not only do IC architectures pose a problem for Darwinism, but more so do fault-tolerant architectures, especially when a fault tolerant architecture is itself composed of several irreducibly complex subsystems! Selection fails to construct fault tolerance because not only do all the parts of subsystem have to be in place for the subsystem to make sense, the existence of the precursors and even functioning subsystems can come at a metabolic cost, especially the large scale fault tolerant systems, making them a liability with respect to immediate fitness.
By definition, Darwinian selection’s lack of foresight will preclude construction of systems that only have meaning through foresight. Construction of a fault tolerant system requires foresight because with respect to immediate fitness, precursors to subsystems are neutral at best, and a liability at worse.
What evidence do we have for this? First some indirect evidence. Consider blind cave fish. An otherwise good visual system that would be beneficial in certain contexts is lost due to the expediency and lack of foresight of natural selection. Selection has no foresight to realize that one day the blind cave fish might be in an environment where its eyes would be useful. Instead, selection eliminates vision rather than preserving it. If selection destroys existing function, how much more will it fail to construct function in the first place, especially when it is not immediately essential and even disadvantageous with respect to immediate fitness.
Blind cavefish provide an illustration of how selection gets rid of function it does not immediately need, and thus how much more will it fail to have foresight to construct something potentially useful. Blind, mindless watchmakers do not plan for future challenges and opportunities, only MINDs do that…
But do we have some direct evidence for the difficulty of selection creating fault tolerance? From Reductive Evolution:
“Interestingly, some species have the ability to regenerate appendages, while even fairly closely related species do not,” Poss added. “This leads us to believe that during the course of evolution, regeneration is something that has been lost by some species, rather than an ability that has been gained by other species. The key is to find a way to ‘turn on’ this regenerative ability.”
If selection has problems preserving fault tolerance, why should it construct it in the first place?
Based on these considerations, I suggest a testable hypotheses: on average real evolution will slowly degrade fault tolerant systems in biological organisms. This can be tested by observing the accumulation of Single Nucleotide Polymorphisms (SNPS) of systems that are not immediately needed, but functional in certain rare contexts. Zebra fish with regenerative capabilities might be a good candidate for such investigations, but I’m sure there are many other candidates…
I demonstrated the fallacy of survival of the fittest in Death of the Fittest, and that real evolution is where function is lost over time, not gained. Fault tolerance is a greater foe to Darwinism than IC, and can lead to further tests of the death-of-the-fittest hypothesis.
In sum, there are three kinds of parts:
1. parts of Irreducibly Complex systems, removal of any of them results in failure
2. parts of Fault Tolerant systems, removal or malfunction of some of the parts does not result in loss of immediate function but reduces the probability of continued function in presence of continued removal or failure of parts
3. useless parts or even parts that are a liability which serve no purpose for the benefit of the organism which can be removed
[Oddly #3 can also pose a challenge for Darwinism in certain contexts (like non-functional convergences), but maybe more on that in a separate discussion on “useless” design…]
IC poses a challenge for Darwinism, and fault tolerance poses an even greater challenge, particularly if the fault tolerant system is composed of irreducibly complex subsystems.
Michael Behe for putting forward the concept of IC on which the further concept of fault tolerance can be constructed.